EPA-600/2-75-016

August 1975         Environmental Protection Technology Series
                       INCINERATOR OVERFIRE
                    MIXING  DEMONSTRATION
                                   U.S. Environmental Protection Agency
                                   Office of Research and Development
                                        Washington, D.C. 20460

-------
                                    EPA-600/2-75-016
  INCINERATOR  OVERFIRE
MIXING  DEMONSTRATION
                  by

   T. J. Lamb, R. H. Stephens, C. M. Mohr,
   P. C. Levins, L. K. Fox, and A. F. Sarofim

           Arthur D. Little, Inc.
              20 Acorn Park
       Cambridge, Massachusetts 02140
          Contract No. 68-02-0204
           ROAP No. 21AUZ-015
        Program Element No. 1AB015

           EPA Project Officers:

            James D. Kilgroe
  Industrial Environmental Research Laboratory
  Research Triangle Park, North Carolina 27711

                 and
           Donald A. Oberacker

  Municipal Environmental Research Laboratory
           Cincinnati, Ohio 45268

              Prepared for
  U.S. ENVIRONMENTAL PROTECTION AGENCY
      Office  of Research and Development
         Washington, D.C.  20460

              August 1975

-------
                       TABLE OF CONTENTS

                                                            Page

List of Figures                                                     v
List of Tables                                                     vii
Nomenclature                                                    ix

SECTION I   -  CONCLUSIONS                                       1

    REDUCTION OF COMBUSTIBLE EMISSIONS                           1
    JET MIXING                                                  1
    FURNACE FLUID FLOW                                          2
    BED BURNING                                                 2

SECTION II  -  RECOMMENDATIONS                                  3

SECTION III -  INTRODUCTION                                      5

    BACKGROUND                                                 5
    OBJECTIVES                                                  6
    APPROACH AND SCOPE                                          7

SECTION IV -  SUMMARY                                          9

    TESTING PROGRAM                                            9
    VERIFICATION OF MODELS                                      9
    STRATEGY FOR EMISSION CONTROL                               11
    TEMPERATURE CONTROL                                       13
    DESIGN GUIDELINES                                            14

SECTION V  -  CONCEPTS IN INCINERATION                            17

    STRATEGY FOR EMISSION CONTROL                               17
    EFFECT OF MIXING                                            18

SECTION VI -  EXPERIMENTAL PROGRAM                             45

    TEST INCINERATOR                                            45
    INCINERATOR MODIFICATIONS                                   49
    CONDUCT OF TESTS                                            50
    ANALYTICAL EQUIPMENT                                       59
    DATA REDUCTION                                              69
                               111

-------
                   TABLE OF CONTENTS (Continued)
SECTION VII - TEST RESULTS

    MACROSCALE MIXING MODEL
    FLU ID-FLOW PATTERNS
    MICROSCALE MIXING MODEL
    RESULTS OF PARTICULATE TESTS

SECTION VIM- DESIGN GUIDELINES

    TEMPERATURE
    FURNACE DESIGN
    UNDERFIRE AIR DISTRIBUTION AND FLOW RATE
    OVERFIRE AIR JETS
APPENDIX A
APPENDIX B

APPENDIX C
APPENDIX D
APPENDIX E
APPENDIX F
APPENDIX G
TEST MEASUREMENT AVERAGES
CALCULATION OF UNMIXEDNESS FACTOR OF REFUSE
BEDOFFGAS
MATHEMATICAL CHARACTERIZATION OF QUENCHING
BASIC PRINCIPLES OF COMBUSTION
CHARACTERIZATION OF THE BED-BURNING PROCESS
DERIVATION OF THE UNMIXEDNESS FACTOR
CONVERSION FACTORS FOR SI UNITS
Page

 75

 75
 76
 77
 87

 93

 93
 93
 95
 98

103

109
113
121
131
145
147
                               IV

-------
                               LIST OF FIGURES
Figure No.

   1
   2
   3
   4
   5
   7
   8
   9
  10

  11
  12

  13

  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
Emission Factors vs.Temperature
Heterogeneous Character of Flue Gases
The Unmixedness Factor  ^/c^ in Turbulent Flames
Concentration Fluctuations
Effect of Stoichiometry on Emissions for a Fixed Turbulence
Level
Effect of Turbulence Level on Emissions for a Fixed
Stoichiometry
Overall Effect of Micromixing on Emissions
Offgas Composition of C0/C02
Theoretical Relationship
Dependence of Emission Factor on Temperature and Unmixedness
Factor
Theoretical Relationship
Dependence of Emission Factor on Temperature and Unmixedness
Factor (Revised  Plot of Figure 10)
Cross Section of Newton 500-TPD  Incinerator Showing
Sampling Locations
Overall Emissions Measurements Approach
Sampling System
Gas Analysis System
Static Pressure and Velocity (AP) System
Water-Cooled Sampling Probe
Thermocouple Radiation Shield
Gas Sampling Lines, Pumps and Meters
Recording Temperature and Pressure Systems
NDIR and FID Analyzers and Recorders
Gas Chromatograph and NDIR Analyzers
C02 vs. Temperature
Penetration vs. Jet Velocity
Confirmation of Bernoulli's Equation
(1-x) vs. (0t)b
Page

 12
 24
 25
 26

 28

 28
 30
 32
 34

 35
 36

 41

 46
 56
 56
 58
 60
 62
 63
 65
 66
 67
 68
 74
 76
 78
 86

-------
                                LIST OF TABLES

Table No.                                                                      Page

    1         Composition of Bed Offgas                                          32
    2         Summary of Calculated Terms                                       43
    3         Velocity and Static Pressure Measurement Locations                    59
    4         Carbon Distribution Profile                                          70
    5         Hydrogen Balance for Typical Run                                   72
    6         Oxygen Balance in Typical Run                                      72
    7         Summary of Unmixedness Factors                                   8^
    8         Summary of Test Results Over Bed                                   82
    9         Comparison of Estimates for (1-x)                                    8"*
  10         Calculation Results Using Different Values for x 	                     °*
  11         Relative Agreement Between L/C**)  t AND (v/c^)taic                  86
  12         Sampling Points Per Radius        eXP                                88
  13         Gaseous Pollutant Emissions                                         88
  14         Stack Particulate Data                                              89
  15         Particulate Analyses                                                89
  16         Trace Metals Analysis of Filter Catch                                 ^
  17         Trace Metals Analysis of Probe and Cyclone Catch                      "'
  18         Excess Air Requirements for Efficient Combustion                     9^
  19         Theoretical Air Requirements of Refuse and Other Fuels                ^6
                                        vn

-------
                            NOMENCLATURE
A             Area
Ac            Area of crossflow, Eqs. (5) and (14)
Af            Area of flow above jets
A,             Area of overfire jet
A0            Refuse bed area
C             Concentration of off gas expressed as mol/volume or as mol/fraction
C             Mean concentration
C             Concentration of off gas at onset of quenching
Cs             Concentration of off gas in stoichiometric mixture
CT            Total molar concentration, mols/volume e.g. (P/RT)
Cp            Heat capacity
Cp   Cp       Heat capacity of combustion products of stoichiometric mixture, of excess air
  s,   xs
c'             Concentration fluctuations about the mean
  c'2         Unmixedness factor (root mean square of c')
         (v c^bo      Unmixedness factor for tests without jets with breech, over bed
(v/c'2);   (\/c'2):        Unmixedness factor for tests with jets with breech, over bed
      J>       Jo
Du            Diffusivity  of fuel in air, gm/cm2 /sec, etc.
dj             Diameter of jet
Ea Eaf Ear    Activation  energy, for forward reaction, for reverse reaction
f(0)          VC^/AC defined by Figure 3
g             Acceleration of gravity
gc             Conversion factor of weight to mass
HHV         Higher heating value
K             Equilibrium constant
K^,           Equilibrium constant for water-gas shift
K E          Kinetic energy
k             Constant or thermal conductivity (App. C)
k3            Chemical kinetic constant of reaction (3), for example
L             Characteristic dimension of system
Lp            Penetration distance of jet
m             Mass
                                    IX

-------
                      NOMENCLATURE (Continued)

N  -          Number of jets
n             Moles per mole of off gas
nc            Moles of combustion products per mole of off gas
no            Moles of oxygen required to bar one mole of off gas
nH  o>nN  > nco        Moles of water vapor per mole of off gas, of nitrogen, of C02, etc.
ns            Total moles in stoichiometric mixture containing one mole of off gas
nT            Total moles in mixture containing one mole of off gas
nxs           Moles of excess air (not oxygen) per mole of off gas
(O2 Req)b, (O2 Req)bo  Oxygen required to complete combustion of off gas without jets in
                       breech, initial
(O2 Req)j              Oxygen required to complete combustion of off gas in breech'
                       with jets on
(O2 Req)jo             Oxygen required to complete combustion of off gas at top of
                       bed with jets on

P             Pressure
Pa            Partial pressure of component a, for example
P0- P          Differential pressure between grate and breech

Q            Flow rate, volume/time
Qc            Flow rate of cross flow to jet axis
Qf            Flow rate of gases after jets or exit the breech
Q.            Flow rate of jets
Q0            Flow rate of off gas from refuse bed

R            Ideal gas constant
S             Jet spacing parameter
T             Temperature
T             Adiabatic flame temperature
 3
Tc            Temperature of gas in cold zone of incinerator
TH            Temperature of gas in hot zone of incinerator
T0            Ambient temperature
T-            Jet temperature
Tc            Temperature of crossflow gases
Tf            Temperature of gases after jets

-------
                       NOMENCLATURE (Continued)

t             Time
u             Velocity
ur ,u          Cold zone velocity, initial value
 *-  co
uc            Crossflow velocity (before jets)
uf            Furnace velocity (after jets)
UH UH        Hot zone velocity, initial value
Uj            Jet velocity
u             Initial velocity
 o
V             Volume of gas
Vf            Furnace volume

x             Moles fraction of off gas in mixture of off gas plus jet air

y             Mole fraction of off gas
yc            Mole fraction of off gas in crossflow (before jets) gases
yf            Mole fraction of off gas in furnace (after jets) gases
ys            Mole fraction of off gas in stoichiometric mixture
Ayb          (y - ys) evaluated for base tests (no jets)
Ay.           (y - yj evaluated for jet tests
   j                 »
z             Distance
zo            Height at top of refuse bed
zc            Height to cold zone in breech
ZH            Height to hot zone in breech
z             Distance from initial surface to quench plane (App. C)
zs            Distance from initial surface to stoichiometric plane (App. C)
z'            Distance from initial surface to zero-oxygen plane (App. C)
                                     XI

-------
Greek Symbols
       a         thermal diffusivity or constant [Eq. (F-4)]
       (3         mixing energy constant [Eq. (22)]
       7         ratio of furnace height to width
       e         rate of input of mixing energy per mass of system
       77         ratio of furnace length to width, also constant [Eq. (C-12)]
       p         density
       pc        density of cold zone gases
       PH        density of hot zone gases
       pc        density of crossflow gases
       pf        density of furnace gases (after jet)
       Pj         density of jet air
       T         residence time
       0         available oxygen/oxygen required to complete combustion
       0b ^b     ^ f°r base tests (no jets), initial value at top of refuse bed
       0 0      0 for jet tests, initial value at top of refuse bed
        j.  JG
Common Subscripts
       b         baseline (no jets)
       C         cold zone
       c         crossflow (before jets)
       f         furnace (after jets)
       H         hot zone
       j          Jet
       o         initial, usually refers to top of refuse bed
       q         quenched conditions
       s         stoichiometric conditions
Gas Compositions
(% CO2)         percent CO2  in gas mixture
(% CO2 )b         percent CO2  without jets on
(% CO2 )j         percent CO2  with jets on
 % CO2 ls         percent CO2  in stoichiometric mixture
CO              carbon monoxide (ppm)
CO|S             carbon monoxide in stoichiometric mixture with defined CO/CO2
% O2             percent oxygen
                                    Xll

-------
                                    SECTION I

                                  CONCLUSIONS

     Overfired air jets have been used in coal stokers to induce turbulence  for increased
combustion efficiency and reduced emissions of combustible pollutants such as CO, H2,
hydrocarbons, and soot. A mixing system of this type was installed in an existing municipal
incinerator under Contract 68-02-0204 with the Environmental Protection Agency. Tests
were conducted to determine the effects of and design criteria for mixing jets. As a result of
the testing program and subsequent analyses, Arthur D. Little, Inc., reached the following
conclusions:

REDUCTION OF COMBUSTIBLE EMISSIONS

     1.   Combustible emissions can occur either by quenching where the temperature
         of the  gases is  too low to sustain combustion, or  by a mixing limitation
         where  the  gases are not completely  mixed with the available  oxygen, and
         thus will not burn out in the furnace within the residence time. The former
         becomes significant when gas temperatures are below about  1400°F; the
         latter is significant when gas temperatures are above about  1800°F. By
         adjusting air input to  maintain furnace temperature within the range of
          1400 to 1800°F, both quenching and mixing limitation were avoided.

     2.   Overfire jets reduced combustible emissions resulting from the mixing limita-
         tion. This reduction is attributed to the greater availability of oxygen  rather
         than to large increases in mixing intensity.

         A theory  based on  statistical descriptions of the  turbulent mixing was
         developed  and verified by test data. This theory  provides a quantitative
         expression  for the  heretofore  qualitative concepts of the three T's - time,
         temperature, and turbulence. This theory can  be used to show under what
         conditions  combustible emissions caused by  inadequate  mixing can be
         reduced by increasing mixing intensity through the use of sidewall jets or by
         increasing furnace residence time.

     3.   Overfire jets  were ineffective in  reducing  combustible emissions due to
         quenching. These  emissions are believed  to be more amenable to control
         through proper furnace design and underfire air control than through mix-
         ing.  Unfortunately, the test incinerator was not sufficiently versatile to allow
         us to explore this emission cause in depth.

JET MIXING

     1.    Test results showed that the penetration of a sidewall jet into the gases from
         the refuse  bed can  be  predicted using the design equations presented  in this
         report.
                                          1

-------
     2.   Sidewall overfire jets  were found  to  be very effective in controlling the
         temperature of the overfire region.

     3.   The jet test configurations, such as opposed jets, interlaced jets, or jets on
         one side, were  shown to be equally effective for attaining adequate uniform
         mixing and temperature control.

     4.   While the jets did not have a significant effect on increasing  the mixing
         intensity in the test furnace, extrapolations of the theory, which was verified
         in the test furnace, suggest that such an effect could be significant in smaller
         furnaces.

FURNACE FLUID FLOW

     1.   The characteristics of the stratified layers, including their acceleration due to
         buoyancy forces, can be predicted from  the geometry  of the furnaces and
         temperatures along the refuse bed surface.

     2.   When the gases exit directly over the cold end  of the furnace (the burnout
         grate) so that no forced  mixing of the hot gases and cold gases occurs, then
         the  gases  will  stratify, forming stable layers. The colder gases  are on  the
         bottom and quench combustion, while the hotter gases are on top and are
         often oxygen-deficient.  Both  conditions  can  result in high combustible
         emissions.

BED BURNING

     1.   The processes of drying, ignition, pyrolysis, combustion, and burnout occur
         simultaneously throughout the refuse bed as witnessed by a constancy of the
         ratio H2 O/CO2 throughout the  furnace.

     2.   The composition of gases  both leaving the refuse bed  and also in oxygen-
         deficient (unmixed) pockets is controlled by the water/gas shift equilibrium,
         as confirmed by test data in both the breech  of  the test furnace and also
         directly over the refuse bed.

     3.   In a furnace  that is unsealed, or which has a large amount of uncontrolled
         overfire air, the oxygen required for combustion is entrained into the refuse
         bed from  the overfire air space. Even when the underfire air is considerably
         less  than stoichiometric, the burning rate of refuse tends to be constant at
         approximately 60 lb/hr/ft2.

-------
                                  SECTION II

                             RECOMMENDATIONS

     Based on  the data collected in this experimental program, Arthur D. Little, Inc.,
recommends the following:

     1.   The use of overfire air jets to  reduce combustible emissions from existing
         municipal incinerators.

     2.   The use of overfire jets in small furnaces to increase mixing intensity. Design
         criteria presented in this report should be utilized.

     3.   The realization of uniform mixing throughout the furnace through proper
         control of underfire air and certain  changes  in furnace configuration. For
         example:

         a.   To achieve uniform  mixing with  underfire  air, a high-pressure drop
              grate  should  be used, and the furnace should  be sealed to prevent any
              leakage which would tend to negate the effect of controlled air rates.

         b.   The furnace  should be designed so that hot and cold gases are forced to
              mix before  exiting the  furnace. This represents  the most  effective
              technique for attaining a uniform front-to-back mixing.

         c.   Overfire jets should be designed to effect uniform  side-to-side mixing,
              as shown in this report.

-------
                                    SECTION III

                                  INTRODUCTION

BACKGROUND

     The  emissions from municipal incinerators have been the subject of  considerable
discussion over the past decade in conjunction with growing public awareness of the adverse
effects of air pollution.  The incinerator emission receiving the most attention  to date is fly
ash,  a  particulate  emission having a  characteristic  dimension  between 0.1  and  100
microns (M). Approximately 70% (by weight)  of  fly ash  has a  characteristic dimension
greater than 5ju and is therefore easily removed using relatively simple devices. The smaller
particles are much more difficult to remove, however, but in recent years several advances
have been made in the state-of-the-art of small particulate removal. The development of
such devices as electrostatic  precipitators  and high-energy wet scrubbers — both high-
efficiency collection devices — has facilitated the removal of particles down to 0.5/u so that
99% of the particulate emission  can be controlled. As municipal incinerators across the
country incorporate these types of emission controls, the prevalence of fly ash as a pollutant
will diminish.

     On the other hand, in its report [ 1 ] "Systems Study of Air Pollution from Municipal
Incinerators" (Contract  CPA-22-69-23 and dated March 1970). Arthur D. Little, Inc. (ADL)
stated  that  the municipal incinerator was  a  large source  of such pollutants  as carbon
monoxide, hydrocarbons and soot (small carbon particles). Unfortunately, these  emissions
are not so easily controlled as is the  fly ash. The soot particles are so small, in fact,  that a
significant fraction will pass through even the highest efficiency control devices. The gaseous
emissions  are relatively unaffected by particulate control  devices. However, these  emissions
are all products of incomplete combustion and, therefore, can be controlled by improving
combustion efficiency rather than by the use of expensive air pollution control devices. In
its report  ADL recommended that methods be developed to  improve the combustion
efficiency of municipal incinerators to preclude incomplete combustion.

     Designers  have  long recognized that  combustible  emissions  could  be reduced by
increasing the combustion efficiency of a process. In fact, they have taken several steps to
maximize residence time and increase the amount  of mixing within the furnace,  but most
current design  practice is based  upon empirical rules or qualitative guidelines.  Unfortu-
nately, many accepted engineering principles of combustion have  not as yet been applied to
incinerators.

     In February 1972, ADL published another report entitled, "Incinerator Overfire Mixing
Study" (Contract EHSD 71-6) [2] .This study,which reviewed the present understanding of
the combustion process in  solid  fuel equipment and municipal  incinerators,  postulated a
technique for improving combustion efficiency in which jets would be used to promote
mixing. A test program was proposed to demonstrate the use of  overfire air and steam jets
to reduce the combustible pollutants being emitted from an existing municipal incinerator.

-------
     In this previous study, ADL postulated that combustion in a municipal incinerator
occurred in three distinct zones — an ignition-limited zone, a combustion-limited zone, and
a fuel-limited (burnout) zone.  Under conditions normally found in incinerators, the oxygen
concentrations  in  the first  and last zones  are high, reflecting the  low oxygen  demand
required to sustain combustion. The central zone is oxygen-limited and has an extremely
low  residual  oxygen concentration.  Similarly,  the temperature in  the  central zone  is
considerably higher than in either of the other two zones and  the gas emanating from this
zone  has  a much  greater concentration  of  CO and other combustible emissions. Unfortu-
nately, the oxygen-rich zones do  not mix well with the oxygen-deficient central zone. The
model developed to determine gas-flow patterns within the incinerator indicated the zones
remain  stratified long after the  combustion process had  been quenched. We, therefore,
suggested  two approaches to eliminate the emissions resulting from incomplete combustion:

     1.    Distribution of underfire air along the  grates in  proportion to the oxygen
          demand, and

     2.    The use of air and/or steam jets to induce turbulence, thus promoting greater
          combustion efficiency.

     To verify  the model for  burning, mixing, and stratification and to demonstrate the
effectiveness of overfire mixing to reduce combustible emissions, EPA awarded a contract to
ADL to test an existing municipal incinerator and to develop design and operating guidelines
which could be used for improving the combustion efficiency of other incinerators.

OBJECTIVES

     Compared  to other  industrial sources  of pollutants,  the municipal  incinerator is a
low-budget piece of equipment, often poorly operated or maintained. There has been little
incentive for either sophisticated testing or  development work on the part of either the
municipality (the owner)  or the  engineer (designer). For that  reason very  few meaningful
data are available.

     To extend  the state-of-the-art, this program had several independent objectives. They
were to:

     1.    Develop  a  basic understanding of  the incineration process by verifying the
          three models presented in the previous report and extending these concepts
          as far  as possible to  describe completely  the combustion of refuse within an
          incinerator;

     2.    Demonstrate the use of overfire jets on an existing municipal incinerator of
          common design and showing their effect on combustible emissions;

-------
     3.   Develop guidelines for design and operation of the jet systems, because much
         of the work is of a theoretical nature extending beyond the normal interests
         of both the designer and the operator; and

     4.   Supply these data insofar as possible, because - independent of the focus of
         the  other  objectives — there is an extreme  lack  of basic operating and
         emission data on incinerators.

APPROACH AND SCOPE

     In developing design data for the use of overfire jets  on municipal incinerators, we tried
to strike a balance between  both the theoretical  and the experimental approaches that are
required  to satisfy the specific objectives of this  program.  Although we only tested a single
incinerator, our intent was to generalize the test results to other furnaces.

     Full characterization  of  the  incinerator  would  involve a  complete input/output
analysis.  Since the principal emphasis in the program was on  combustible emissions, the
major data-collection effort  was restricted to this area. Limited data on material input were
obtained to guide the overall data  analysis. The information obtained during each of the
tests included:

     •   Input: refuse tonnage, grate  speeds and air flow rate;
     •   Output: temperature, flow (velocity), and composition.

     Gaseous  composition data were required  to  reflect both the overall  combustion
performance  of the  incinerator and the variation in combustible emissions. The following
species were routinely analyzed in the  breaching region (and over the burning bed during the
study of that region):

     Overall composition: O2, CO2, N2, and H2 O;

     Combustible emissions: CO, H2,  and THC (total gaseous hydrocarbons).

     In  addition, a carbon monoxide  analyzer was run continuously in the duct leading to
the base  of the stack to reflect the long-term trend in changes of emissions.

     The key  factor in  the analytical testing program  was  the development of a  high-
temperature testing probe to sample in the 2000° F (or greater) conditions found both in the
breech and over  the burning bed.  Continued  service  from  a  test probe  under these
conditions  for more than one  week of testing proved to be  very difficult, but because of
modification in the testing procedure and improvements in the probe design, we believe that
the technique developed in this program will be  generally  applicable to all similar instances
of emission testing in high-temperature environments.

     In  conducting  this program, the initial series  of tests  (the baseline tests) was  used
primarily to verify the analytical approaches and  to ensure that the test data were consistent

-------
internally and with the physical concepts of incineration. With each series of tests, material
and energy balances were completed and the results were analyzed to indicate potential
weaknesses in the test data or to point out areas where present understanding of combustion
could not correlate the experimental results. Because of this procedure, there were changes
made in  the  analytical  approach with corresponding increases in the accuracy of the
experimental data. The material balances for the final jet tests indicated the magnitude of
error in the analytical data to be less than 10%.

     Upon verification of the initial testing data, it became apparent that many of the
theories  used  to model  the combustion were  not  wholly adequate with respect to the
characterization  of mixing  and  combustible  emission burnout.  Additional work in the
modeling of combustion was carried out to  augment the models presented in our previous
report. The emphasis was on increasing the sophistication of the models so that the physical
phenomena  could  be more  readily deduced from the test data.  The analysis of the bed
model took on much less importance as the work progressed. The bed model is introduced
only in the Summary to the report. The detailed discussion of the bed model is contained in
Appendix E.

     We  did not intend for this report  to be an exhaustive compendium of combustion
mechanics, but only that the principles and experimental results pertinent to the burnout of
combustible emission be fully explored and documented. We have incorporated the test data
in as complete a form as  possible (Appendix A), and we feel it will prove to be a valuable
resource  to those workers in the field who would like to extend this work still further, or
who may require data on existing furnaces for experimental support of their own theories.

-------
                                    SECTION IV

                                    SUMMARY

     The primary  objective of this program was  the reduction of combustible emissions
from existing municipal incinerators through the use of overfire air jets to induce mixing.
The objectives cited in the Introduction of this report represent the various aspects required
to study this particular problem adequately.

TESTING PROGRAM

     To provide the data required to verify the burning and mixing models developed in the
Incinerator Overfire Mixing Study (2), we planned a test program at an existing municipal
incinerator. We chose the  incinerator of the City of Newton, Massachusetts,  as the test
vehicle, because it  is similar in design to many incinerators in the United States, it could be
easily retrofitted with an air jet system,  and the City of Newton was kind enough to make
the facility available.

     In the initial  tests, we recorded  composition, temperature, and velocity data in the
breech of the furnace to  check  out  the  analytical procedures and provide information
needed to design the jet system.

     Following the procedure outlined in the previous study, we designed and installed an
overfire jet mixing system. The system consisted of ten 4-inch jets on each side  of the
furnace, a-forced draft fan, and the ducting and instrumentation necessary to control the
system.

     The test program included tests both  in the breech of the incinerator and also over the
refuse  bed. The breech tests  were required  to verify the fluid flow model  and also to
confirm or deny the effectiveness of overfire mixing jets in reducing combustible emissions.
The bed data were  necessary for verifying the bed burning model.

VERIFICATION OF MODELS

     In the previous study  (2), ADL postulated three models necessary to describe incinera-
tion:  l)a bed-burning model, 2) a  fluid-flow model, and 3) a jet-mixing model. All three
were verified either as postulated or with the refinements added as a result of the  testing
work. Each model  is summarized below.

Bed-Burning Model

     The bed-burning model was originally proposed to describe the overbed gas composi-
tion and also the burning rates for refuse beds. The model was previously postulated on the
basis of separate drying, ignition, pyrolysis, combustion and burnout planes, but the results

-------
of the testing program showed that the processes occurred simultaneously throughout the
refuse beds. However, this fact does not alter the concept that the gases leaving the refuse
bed in the absence of oxygen are controlled by the water gas shift equilibrium. In support of
this, gas samples taken directly over the refuse bed contained concentrations of CO, CO2,
H2, and H2 O in direct proportion to what would have been predicted from the equilibrium
relation. However, each sample also contained a significant amount of oxygen, although the
previous model was based upon the existence of a plane of zero oxygen within the refuse bed.
This particular phenomenon strongly suggested that  the combustible gases were not mixed
with oxygen  and remained unmixed even after the gases had  left the furnace. As a result a
fourth model — the  microscale  mixing theory — was  deemed necessary to describe com-
pletely the combustion phenomenon that was occurring within  the furnace volume. This
model is described in a later section of this summary and also in Section V.

     Based upon the empirically determined amounts of oxygen that can be entrained in the
refuse bed and  also upon  the  water gas shift gas equilibrium,  we were  able to develop
estimates  for overbed composition of gases. This information was required in the model
developed for microscale combustion.

     An additional  ramification  of the oxygen  entrainment was the fact that the bed-
burning rates tended  to remain constant at approximately 60 lb/hr/ft2. Surprisingly enough,
the bed-burning rate  was not very sensitive to the  underfire air rate in that the air deficiency
was supplied from the overfire region. This would not be the case either when the amount
of oxygen was very limited, or when greater than stoichiometric amounts of underfire air
were present.

Fluid Flow Model

     The  acceleration of gases away from  the refuse bed can be mathematically  charac-
terized using Bernoulli's equation:
                                      2g(z-zo)
(1)
In a furnace that is completely sealed, the appropriate Tc is the temperature in the cold
region of the furnace, most likely the temperature over the burnout grate. In an unsealed
furnace the  appropriate Tc  is the ambient air temperature. Because  of the large amount of
leakage that was inherent in all of the test runs at the Newton incinerator, it was obvious
that the furnace was unsealed. Equation (1) was verified for a Tc of approximately 70°F.

Jet Mixing Model

     In a previous study  (2), ADL postulated that jets could be used to mix cold regions,
such as the region over the burnout grate, with the hotter regions of the  furnace  so that the
overall temperature of the furnace could be controlled. To design such a jet it was necessary
                                         10

-------
to arrive at  a relationship between the jet penetration distance and the parameters of the
system. The jet tests were able to verify the design equation suggested by Ivanov (3):
                                                                                (2)
In designing a jet system, the designer has the freedom to adjust the jet diameter location
and the jet velocity, with the only constraint being that the jet system not be allowed to
deliver an amount of air to the furnace that would quench combustion. The other variable
in the jet design equation is uc , the cross-flow velocity for the gases rising off the refuse bed
perpendicular to the row of the jets. This is determined from Eq.(l) above.

STRATEGY FOR EMISSION CONTROL

Origin of Combustible Emissions

     Two phenomena which result in combustible emissions were identified in the course of
this study. The  experimentally determined CO emission factor for the furnace is a function
of local  furnace temperature as shown in Figure 1. In regions of the furnace where the
temperature was below approximately  1 100°F, emissions occurred because the reaction had
been quenched, i.e., it was too cold to sustain combustion.

     On the other hand, when the temperature was high (above 1800°F), the corresponding
amount of excess air available was very low. In the case of perfect mixing, this has no effect,
but for incineration, we postulated that the gases leaving the furnace have a heterogeneous
composition consisting of complete separate pockets of oxygen containing combustion
products and "fuel" containing combustion products. At high  temperatures "fuel" and  air
do not co-exist  in a pocket  for long, because the  combustion  reactions occur too rapidly.
When the amount of available oxygen  decreases (as temperature increases), the probability
of finding a pocket of "fuel"  that has not mixed with available oxygen  increases. This
mixing limitation, as indicated by higher CO emissions, was observed for temperatures above
about 1800°F.

Reduction in Emissions

     The key to the elimination of combustible emissions is enhancement of combustion
efficiency in the overfire region of the furnace. For a long time designers have recognized
the importance of the three T's - time, temperature, and turbulence. In this report we have
gone one step farther in developing the necessary mathematical models for describing the
various phases of the  operation  so that the interrelationships between the three T's can be
quantitatively described.
                                        11

-------
t-o
                       7000
                       6000
                       5000
                  -    4000
                  @
                  8
                  oc    3000
                  O

                       2000
                       1000
                                             t             t
  Quench
  Regime.

                 1
                 \
                  \
                  \
                   \
                 •   \


                                                        .Mixing
                                                         Limitation
                                                         Zone
                                                                                    X*      •
                                   1     1      1      1      t*   1      1      !•    1      1      1     1      1      1
                            700
1000
                                                                         1500

                                                               TEMPERATURE (°F)
                                                        2000        2200
                                                    Figure 1     Emission factors vs. temperature.

-------
TEMPERATURE CONTROL

     The data of Figure 1 clearly show the higher CO emission at temperatures above about
1800°F or below about 1400°F. We believe the high-temperature emissions are caused by a
mixing limitation resulting from a relatively low amount of available oxygen (excess air). As
more excess air is added (consequently lowering gas temperature),  the mixing constraint is
eased and the high CO emissions are reduced. In the case of high-temperature emissions, jets
can be used to add a  sufficient amount of overfire air to  control the excess air within the
given range. In designing such a system, it is important to obtain a uniform "macroscale"
mixing throughout the furnace. The specific design criteria  are discussed later in this section.

     Using this type of control, the low levels of combustible emissions (as obtained in
Figure 1) can be achieved.

Mixing

     As pointed out  above,  it  is important to  maintain a uniform macroscale mixing
throughout  the furnace, primarily to  avoid hot  and cold spots, i.e., to maintain the gas
within the 1400 to  1800°F range. However, in terms of burnout of combustible emissions
one  must also consider the problem of microscale  mixing in addition  to the  macroscale
considerations of uniformity. The microscale problem is concerned with the mixing within a
local region of the furnace, or along  a given streamline, to bring the  combustibles and
available oxygen into intimate contact with  each other so that the desired burnout can
occur.

     To describe this macroscale mixing, it was necessary to  introduce two well known
concepts that have not been previously applied to incinerator designs. These are:

     (1)   The use of the unmixedness factor to characterize the level of emissions; and

     (2)   The use  of  statistically based mixing theories to  describe the relationship
          between  the degree  of mixing and  the time required to burn out a given
          pocket of combustibles.

These  concepts  are discussed  in  detail in  Section V. A qualitative discussion  of these
concepts will be included here.

     •    Unmixedness

          For complete combustion, mixing on a molecular level is required. A single
          sample containing both oxygen and combustibles reflects alternate sampling
          from oxygen-rich and fuel-rich gas  pockets. A  measure  of unmixedness of
          the  overall  sample can be deduced from  the ratio of the concentration of
          oxygen available in the gas sample to that required to complete the combus-
          tion process using the statistical description of turbulence. This measure [as

                                        13

-------
          suggested by Hottel (4)] reflects both the intensity of turbulence as well as
          the relative concentration of the combustibles and air.

          The emission data shown in Figure 1 correlate with furnace temperature and
          Hottel's (4) unmixedness factor to the extent that the scatter in the data can
          be resolved within about 10%.

     e    Microscale Turbulent Mixing

          The final intimate mixing of oxygen and fuel can be described as a turbulent
          fluctuation decay process analogous to eddy decay  phenomena that have
          been extensively studied in wind tunnel experiments.

          The mathematical relation suggested by Corrsin (5) equates the unmixedness
          factor  (^/c'2)  to  an exponential function of residence  time and mixing
          intensity (turbulence).

     Mixing intensity and residence time are both affected by the furnace design. Since the
unmixedness factor can be related to design parameters, based on Corrsin's equation, and to
the  emission  factor, based  upon Hottel's relation, this  defines the overall quantitative
relation between emissions and the three  T's.  The  data taken  in this program were
sufficiently  accurate to support this conclusion,  but not sufficiently accurate  to  provide
proof of the hypotheses.

DESIGN GUIDELINES

     To  achieve  the desired reduction in combustible emissions,  a proper design must
address both the macroscale and microscale mixing effects. The uniform, or macroscale,
mixing can  be achieved both  through the proper use  of overfire jets and through careful
design of the furnace enclosure, considering both the configuration and furnace dimensions.
The jet design strategy  is based  upon the method of Ivanov (3)  which could be  easily
verified from the test data. Several  tests were run using various jet positions, including
interlaced, opposed, and  one-sided jets, and all were  shown  to be equally effective in
providing uniform macroscopic mixing and temperature control capabilities. The jets proved
to be most advantageous when used to achieve an overall side-to-side mixing and to maintain
the overfire temperature within the proper range.

     Overfire jets have been suggested also  for use  as a means of obtaining front-to-back
mixing in the furnace. However, we believe the same effect  would be more easily obtained
through the proper design  of the furnace  configuration and  proper control  of  the air
distribution  so that quenching zones are prevented from occurring, and so that hotter gases
are mixed with  the colder gases to attain a uniform temperature throughout the furnace.
The latter is  achieved by designing  a furnace with flow patterns so  the  hot gases are
deflected  into the colder gases, or in some other way forced to mix with the colder gases,
rather than allowing stratification to occur, as that demonstrated in the testing program.

                                         14

-------
     In addition, the furnace should be a sealed enclosure so that maximum control of the
air  is maintained through  the elimination of leakage. Moreover, the roof of the furnace
should be fairly close to the refuse bed so that the gases rising from the bed do not have a
chance to accelerate under the influence of buoyancy forces. By limiting the velocity of the
gases, the jets will be able to achieve a much greater control of the overall mixing patterns
within the furnace.

Microscale Effects

     Overfire jets can also be used to increase the mixing intensity of the overfire gases, but
care must be taken to avoid consequential quenching of the combustion reactions. The
primary use of the jets is for the macroscale mixing and temperature control, but if the jets
can be so designed to accomplish  these major objectives and still be capable of adding
sufficient mixing power to the overfire regions to increase the  microscale  turbulence level,
then so much the better. In our experience during the tests, we found that the amount of
power that  would be  required  to attain  a noticeable effect  on mixing intensity  was
extremely large so as to be prohibitive. The mixing power required to achieve a given mixing
intensity  scales as the fifth power of the characteristic dimension of the  furnace. Calcula-
tions show that for the large municipal incinerator the effects of jets on microscale mixing
are only marginal, being effective in some cases and ineffective in other cases. But in smaller
furnaces,  the  power requirements are  greatly reduced so that jets could be significant in
affecting  the  mixing intensities. As a result of the tests, we were able to  develop a design
criterion to determine in which  cases the jets could be used  to attain a mixing intensity
advantage, in addition  to  accomplishing the  macroscale mixing and temperature  control
functions. This is discussed in  Section VIII. The  criterion is based upon the quantitative
relationship between the emission factor and the three T's that was developed during the
course of the program. The  full utilization of this  criterion and of the other  conclusions
derived from  this program will  represent a  major step forward in the  development of
incineration technology within the framework of combustion engineering.
                                         15

-------
                                    SECTION V

                          CONCEPTS IN INCINERATION

     Incineration  has proven very  difficult to describe quantitatively, because  of  the
simultaneous occurrence of a great number of very  complex phenomena. The description of
a combustion  process must  include  fluid mechanics, thermodynamics,  heat and mass
transfer, reaction kinetics, and statistics.  However,  these theoretical principles for the most
part  have been reduced to rules-of-thumb when applied to the design of an incinerator. For
example, designers assume burning rates of 60 lb/hr-ft2  and heat release rates of 25,000
Btu's/hr-ft3, basing their designs on past experience or upon attempts to maximize  the
effects of the three T's: time, temperature, and turbulence.

     In this chapter, we will describe the processes which occur during incineration, using
the mathematical models developed as a part of this  and the previous study  on overfire
mixing  (2). In  addition, we will propose  a fundamental  basis for many of the common
design rules. Required background material related to the basic principles of combustion —
chemistry, kinetics, equilibrium, and the like — is given in Appendix D.

STRATEGY FOR EMISSION  CONTROL

     Two phenomena which result in combustible emissions were identified in the course of
this study.  The  experimentally determined emission factor for the furnace was shown as  a
function of furnace temperature in Figure 1. In regions of the furnace where the tempera-
ture  was below approximately  1100°F,  emissions  occurred because the reaction had been
quenched, i.e., it was too cold to sustain combustion.

     High levels of emissions are also associated with high furnace temperatures. It is  not
uncommon to measure concentrations of CO or H2 on the order of 1000 ppm in flue gases
containing  10% O2 and at 2000°F. These emissions are  the result of incomplete mixing, i.e.,
combustibles and oxygen have not come into intimate contact at any time during the entire
passage through the furnace.

     The key to the reduction of both sources of combustibles emission lies in  proper
control of residence time, furnace temperature, and turbulence. Residence time is a design
parameter of the furnace. During the testing of the existing incinerator this parameter could
not be varied. The dramatic effect of local furnace temperature can be seen in Figure 1. The
third parameter —turbulence — and the related rate of  mixing are  more difficult to  charac-
terize than either time or temperature. However, mixing  is much more controllable  in tests
than  the other  two parameters. For example,  the most direct  approach to controlling
high-temperature emissions is to add more air; this process is controlled by mixing. As
another example, a possible approach  to eliminating cold-temperature emissions is to force
gases emanating from relatively cold areas of the furnace to pass through the hotter gases in
the furnace. This process is also controlled by mixing.
                                        17

-------
EFFECT OF MIXING

     In the test  program, high-velocity overfire  air jets  were installed on an  existing
incinerator so that the effects of overfire mixing could be more fully explored. During the
course of the work two mixing concepts emerged:

     1)  Macroscale mixing - required to obtain uniformity of gases throughout the
         furnace so as to eliminate or minimize the effects of cold or hot spots; and

     2)  Microscale mixing — required to bring molecules of oxygen and combustible
         gases together so that complete burnout of the combustible emissions could
         be obtained.

     These two scales of mixing are  completely different in character and  are discussed
separately.  However, the optimum strategy  for reduction of combustibles is common to
both mixing scales, i.e., supply the available oxygen at selected locations in sufficient excess
to burn out the combustibles in a reasonable time without quenching the reactions.

Macroscale Mixing

     Non-uniform mixing throughout  the furnace is quite common because of the extreme-
ly diverse combustion conditions observed over individual grates. Proper use of air jets firing
across the fuel bed results in two benefits on the macroscale:

     1)  The jets supply additional oxygen  so that the stoichiometry can be  adjusted
         to the optimum level for control of high-temperature emissions; and

     2)  Jets can break up the stratification of the combustion gases so that oxygen-
         deficient (hot) regions and oxygen-rich (cold) regions are mixed.

     To perform the required mixing functions, a  properly designed jet must:

     o   penetrate far enough into the combustion chamber to influence the macro-
         scale flow patterns; and

     o   carry sufficient oxygen to provide for combustion of the fuel components
         present in the gas, but not so much as to quench the reaction mixture.

     Careful design is required to satisfy both criteria.

     Jet Penetration — The design procedure for an overfire jet  system was worked out in
the previous mixing study based on the method of Ivanov [3]. In summary, antappropriately
designed jet is one in which the momentum forces predominate over  buoyancy forces. Using
these criteria and considering the  basic fluid mechanics of turbulent jets, Ivanov presented
an equation for the penetration of the jet as follows:

                                        18

-------
                                                   dJuJ
                                                    u,.
(3)
where
              Lp   =   penetration distance

              p    =   gas density

              dj   =   jet diameter

              u    =   velocity
and subscripts
              j    =   jet
              c    =   crossflow (in furnace).
     Based upon an empirical analysis of jet interactions, Ivanov suggested a jet spacing of
between 3 and 5 jet diameters. Since the flow rate from a given jet can be calculated from
the equation:

                          «-_^L  .  ..     *     LpUc
                                                      1.6
                                                                                  (4)
and each jet covers an area of grate:
                                   Ac = (Sdj) Lp
(5)
where
                      S = the spacing parameter (between 3 and 5),
then the amount of air introduced over a unit area of grate is:
                                   QJ      TT    »c      pr
                               _  _   J	    .  I   ._                     f£\
                               (J. —      —             •  I                         \OJ
                               4J   Ac     4   1.6S   Y PJ                      V '
which is dependent upon only the crossflow velocity, the spacing parameter, and the density
ratio. The crossflow velocity  can be determined from the  model for  furnace  flow as
discussed below.

     Furnace Flow — In most existing municipal incinerators, flames are concentrated in the
initial portions of the furnace and are  considerably diminished toward the end of the grates
where the refuse burns out. Previous models of the fluid-flow patterns in an incinerator have
postulated the existence of stratified zones [2]. Hot furnace gases from the front end of the
                                        19

-------
 furnace tend to rise because of buoyancy and exit the furnaces along the roof of the breech.
 Colder gases from the after sections form a layer exiting along the bottom of the breech.
 The two zones are stable and will not mix. The colder zone is often below the quench
 temperature, resulting in  high  levels of  combustibles. Provisions for mixing  of the two
 stratified zones will create a uniform condition more conducive to combustible burnout.

      In our previous mixing study  [2], we  used a two-zone model to characterize the
 differences between the hot gases and cold gases in the furnace. Bernoulli's equation can be
 applied to both the hot zone and the cold zone, and the two equations can be combined as
 follows:
                          Hot Zone
                                          Cold Zone
                 H
                            PH
                   gr -2g(z-z0)
                                                     (Po - P)
- =(z-zn)
                                                                                (7)
Combining (APH = APC) yields:

                                -1
                                     u' = u*
                                      C    C0
                                                                                (8)
      The hot gases accelerate in traversing from the bed to the breech as a result of the net
 buoyancy force created by the temperature difference between the hot and cold zones. As
 the  temperature  of the hot zone approaches that  of  the  cold  zone, the  amount of
 acceleration is reduced. Note that the cold gases do not accelerate, but leave the furnace at a
 constant velocity. If the furnace is not sealed (as in the case of the Newton incinerator),
 then there is an interaction between the inside and outside of the furnace such that both the
 hot and cold zones  accelerate with the hot zone accelerating faster than the cold zone. The
 equations for flow in this case are based on the ambient temperature T  as  follows:
                                                                0
      UH=UH,
                 Hot Zone
2g(zH-zJ
                                     Cold Zone
•-1
+ 2g(zc-z0)
      -1
                                      (9)
 Combining the two equations and assuming that (ZH — z0) equals (zc — z0) yields:
2g
                                              (z-z0)
                                    /TH-TC\"
                                    \   T0    /
                 (10)
      To digress for the moment, note that the time required for the gas to travel a distance
   along a streamline can be calculated by integrating the equation about the streamline:
                                     • S.
                                   =f  dfi(x,z)/uz
                                                               (11)
                                         20

-------
      The residence time can be calculated from this relation if the path of the gas is known.
 An  analytical solution  to  this problem is quite complex and graphical procedures are
 therefore preferred. In determining the time required to reach a given position over the bed,
 a simplifying assumption that the direction of motion near the bed is always vertical results
 in an analytical solution to the above equation for the hot zone:
      T =
          g(TH  - To)

                                                                                (12)
 Note that if I/Hospital's  rule  is applied in taking the limit as TH  approaches T0 ,  the
 equation reduces to:
                   = (z - zn)/up
                                                                                (13)
 This is a form of the familiar equation:
                                  Time = Distance/Rate

which applies to the cold zone (or any non-accelerating zone) under the assumptions made.
      Stoichiometry - The air rate required to achieve the optimum stoichiometry is not at
 all a function of fluid mechanics, but rather is dependent upon a difference between the
 actual and the optimum concentrations of offgas emanating from the bed. In the section on
 microscale mixing, methods will be discussed for estimating these optimum concentrations.
 Consider the effect of the jet in mixing air with the offgas from the bed. A material balance
 can be made around  the zone  of jet action as shown in the illustration below. A material
 balance, assuming no  change in temperature in going from the bed to the jet level, results in
 the equation:
 where
           uc  =
           u0  =

           Ac  =
                                          = ucAc
crossflow velocity,
initial velocity leaving the bed,
crossflow area at the jet level, and
bed area.
                                                             (14)
                                          21

-------
                        Hot Zone //Cold Zone'

                                                    Qj
                         Refuse
     Since the area of coverage of each jet is fixed by the jet design, according to Eq. (5)
above, then  the amount of gas  coming from the bed that must be stoichiometrically
adjusted  by a given jet can be calculated. Note that if the jets are evenly spaced, then the
jets in the hotter sections  of the furnace interact with  gases from proportionately larger
grate areas than  the jets in  the colder section of the furnace because of the "necking down"
process,  common with  accelerating gas streams.  The volumetric flow into the  zone of
influence of the  jet is therefore equal to uc xA.c . On the  basis of ideal gases, a mass balance
can be made about each jet  zone as shown in Eq. (15) below:
                               jiT.  QciTc   QfiTf
                              7    + ~,   = ~,                                (15)
where QlT = volumetric flow rate evaluated at temperature T.

and subscripts j, c, and f = jet, cross-flow (before jets), and furnace (after jets), respectively.

Finally, a material balance of the offgas leads to the relation:

                          Vc -CTc -QclTc=  yf -CTf -Qf'if                     06)
where     y   =    mol fraction of "offgas," and
          CT  =    molar concentration (CT  = P/RT).

and  the subscripts c = crossflow condition (before jets), f = furnace condition (after the
jets).
*
This equation can be combined with Eq. (15) and the ideal gas law to derive the relation:
                                       <4IT.    / T  \
                          yc- yf  _ —J_!    (A_)                        m}
                          	      O I      \ T   /                        (1 '>
                            Yf         g'lTc    ^TJ  7
                                         22

-------
By rearranging and substituting for Qc the jet flow rate can be determined by:
     This equation differs from the one based on jet penetration criterion Eq. (4), not in the
design variables dj, Lp, and uc, but in the spacing parameter S, the temperature ratio, and
the concentration terms. If the concentration term is determined to achieve the optimum
emission level,  i.e., if Cf is  the  concentration  at  the minimum point of the  U-shaped
emissions curve (Figure 1), then  there is no  assurance that complete penetration will be
achieved. Likewise, if the jet flow rate is designed for complete penetration, there is no
assurance that the optimum stoichiometry above the bed will be attained.  The skill of the
designer is  then  important  in being  able  to balance the  tradeoffs between  complete
penetration and optimum stoichiometry.

     To further specify the design, note  that  Eqs. (4) and (18) could be set equal to each
other and simplified to the relation:
                                            6.4S     *  «'  '                    09)
     If the composition of offgas emanating from the bed could be predicted, then the left
side  of  Eq. (19) could be evaluated. The equation would then provide a basis for deter-
mining S or the number of banks of jets that could be used to introduce all of the required
air. Alternatively, the jet  penetration criteria  could be  used to determine the amount of
overfire air required. Equation (19) then forms the basis for determining the underfire air
distribution necessary to produce the optimum stoichiometry. At the present time,  our level
of understanding does not allow us this additional degree of freedom.

Microscaie Mixing

     The co-existence of combustibles such as CO  or H2 and oxygen at high  temperatures
suggests that the gas has  a heterogeneous consistency,  as shown in Figure 2. The rate of
dissipation and eventual burnout of the entrained pockets of combustibles is controlled by
mixing  at the microscopic level, i.e., microscale mixing. The problem is  to correlate this
mixing mechanism with the reduction of emissions.

     The  characterization  here is adapted  from  turbulent fluid mechanics and mixing
studies  in high-velocity  burners. Two distinct concepts  are  required  to  complete the
characterization:

     1)    The relation of emission factors to fluid mechanics concepts, and
                                         23

-------
     2)   The relation  of fluid mechanics concepts to incinerator design parameters
          (the three T's).
                                                       7
                                                                 COMBUSTIBLE POCKETS
                                      AIR POCKETS
                    Figure 2  Heterogeneous character of flue gases

     Emission Factors and Fluid Mechanics — The mixture of complete and partial combus-
tion products emanating from the refuse bed is a gaseous "fuel," and the mixing process can
be  characterized as  being similar to  the mixing in turbulent flames (as  suggested by
Hawthorne, Wendell, and Hottel [4]). The  characterization which  results in a statistical
description of turbulence can be summarized as follows.

     In turbulent systems, concentration at any point in the system  can be represented by
two components: (1) the time mean value concentration, and (2) the fluctuating component
of the concentration. These are represented by the equation:

                                      C = C + c'                                  (20)

where c' has been shown  to be randomly  distributed about C. The units of C can be either
moles/volume or mole fractions.

     By applying the gaussion distribution to c', Hawthorne et al (4) showed  that the ratio
$, of oxygen in a sample  to the oxygen required to complete the combustion of the sample
could be used  to determine the values of the parameter y/c'2/(C—Cs)* where-y/c^is the
r.m.s. value of c' and (C-CS) is the difference  between the initial "fuel" concentration C
before any combustion occurs and the stoichiometric  concentration  Cs,i.e., that initial
concentration of "fuel"  which would just consume all oxygen  present. Note that  the
variable C represents the concentration  of "offgas" in air and not the concentration of the
'Hawthorne's derivation uses C to represent either mol fraction or molar concentration interchangeably.
 The same convention is used throughout this report where C is often used interchangeably with y, the mol
 fraction of offgas.
                                       24

-------
constituents  making  up the  offgas.  Also,  note that C is not an average value of the
concentration like Cs, but rather an instantaneous value which is a function of time and
position.

     Hawthorne et al (4) further showed that  which can be calculated from the sample
analysis, can be used to determine the nature of the unmixedness factor^/c2^ based upon
the relation shown in Figure 3. The derivation of this relation is in Appendix F. An example
of this calculation is given in Appendix B. The relation will be shown later to correlate quite
well with the emission data for incinerators.
                              s
                                             •0 40  *0
                                                            MO  MO 400
                             ACTUAL OXYGEN IN SAMPLE	
                        02 REQUIRED TO COMPLETE COMBUSTION
             Source:  Ref. (4).

                  Figure 3   The unmixedness factor   y  c'2  in turbulent flames
     The value (AC) can be calculated if the original offgas composition is known. For pure
fuels, the calculation is straightforward;  for incinerators, the composition of offgas ema-
nating from the refuse bed must be assumed. This is examined in a later section of this
report. However, if AC_can be determined, then Figure 3  can be used to determine the
unmixedness factor, VA^27 Hottel (4) has proposed that this factor be used as the character-
ization factor for c' to relate emissions to fluid mechanics.

     To illustrate the relation of this concept to emissions, consider the time distribution of
concentration, shown schematically in Figure 4, where C is the concentration of refuse bed
offgases in air before combustion, Cs is the stoichiometric concentration of offgas in air, and
                                        25

-------
Cq  is the concentration* of the offgas at the onset  of quenching. Concentrations which
exceed  Cs  represent regions which,  if well mixed, would not have sufficient oxygen for
complete  combustion. Any  such regions  surviving until  the gases leave  the furnace will
contribute  to the  total  emission of combustibles.  Likewise,  concentrations below Cq
represent regions where combustion  has been quenched as a result of low temperatures. If
these regions contain combustibles, then they too contribute to emissions.
 u.
 H UJ
 <0 C
 Zc/5
 UJ _1
 O UJ
 I2
MEAN
                                           HIGH-TEMPERATURE EMISSION REGION
                                               LOW-TEMPERATURE EMISSION REGION
                                                 TIME
               'Expressed as fraction reduced to unreacted constituents.
                                   Figure 4   Concentration fluctuations.

     The source of emissions and the unmixedness factor come together when one notes
that the unmixedness factor, v c'2 . is the standard deviation of concentration fluctuations
about the mean concentration. If the distribution is gaussion, as Hawthorne et al (4) suggest,
then C  and vc71 are sufficient to determine the entire distribution. Furthermore, one can
determine Cq and Cs from the composition of the offgas, so that the amount of gases within
the concentration distribution falling outside  the  region from Cq  to Cs can be estimated.
The techniques for control then become much better defined.
     From the analysis of the two emission-producing conditions discussed above, it is clear
that combustible emissions can be minimized or avoided if the concentration fluctuations
*Since concentration and temperature are directly related, a quench temperature can be represented by a
 corresponding concentration.
                                         26

-------
are maintained entirely between Cs and Cq. This can be done either by adjusting the local
air-to-fuel ratio or by increasing the level of turbulence. These two approaches, illustrated
below, in Figures 5 and 6, are the province of microscale mixing.

     With respect  to the stoichiometry, consider the  distribution of fluctuations in fuel
concentrations shown in Figure 5. In distribution A, the mean concentration is close enough
to Cq so that  a large portion of the distribution lies within the quench zone and emissions
resulting from low furnace temperatures are expected. In distribution C, on the other hand,
the mean concentration is close enough to Cs so that a large portion of the distribution has a
concentration  greater than Cs and emissions will result because of the mixing limitation.
Clearly, the emissions can  be minimized, as in the case of distribution B, by adjusting the
mean concentrations so  that the portion of the distribution outside of the range from Cq to
   is minimized.
 •"s
     The above example illustrates that the value for C can be shifted between Cs and Cq
without changing the level of turbulence in the system simply by adjusting the excess air in
the furnace,  i.e., by controlling the furnace temperature. Of the two variables used for
controlling emissions, this one is by far the most useful.

     The effect of  the level  of turbulence is shown in Figure 6. The greater the level of
turbulence, the narrower the frequency distribution becomes, increasing the probability of
operating the furnace with  no concentration  fluctuations outside  the  combustion zone
range. In distribution A, the mixing is poor; it is increasingly better in distributions B and C,
respectively. The limiting case, of course, is that of infinite mixing (distribution D), resulting
in zero emissions for all concentrations between Cq and Cs.

     Fluid Mechanics and the Three T's — A  considerable amount of research has been
devoted  to  the  study of turbulent  mixing,  primarily based on the statistical theory of
turbulence as applied to the aeronautics industry. Correlations between velocity fluctuations
behind a grid in a wind tunnel and stream parameters  have been thoroughly studied.  As a
result,  the decay of eddies  (represented by eddy velocities) with respect to time is both
empirically and theoretically described. This work is reviewed in References 5 through 8.

     Relying upon  the familiar similarities among heat and mass and momentum transfer,
Corrsin (5)  suggested the concentration analogy  to velocity decay, reasoning that fluctua-
tions in concentration and velocity both follow the same equations. The fundamental decay
equation is:

                                 yc^X/cf" = e-#   .                            (21)

The factor 0 is defined by the equation:

                                                                                 (22)
                                          27

-------
      MEAN
                                                  MIXING
                                               LIMITATION
                                                  ZONE
            Q:
Figure 5  Effect of stoichiometry on emissions for a fixed turbulence level.
                                                       MIXING
                                                    LIMITATION
                                                       ZONE
Figure 6    Effect of turbulence level on emissions for a fixed stoichiometry.
                            28

-------
where

     L   =    the characteristic dimension of the system,
     e    =    the rate of input of mixing energy per mass of system, and
     k   =    0.19, as determined from turbulence experiments in wind tunnels [5].

     Since the  term y/c'2  can be related directly to the emission factor of the furnace
(through the stack gas analysis), this equation represents the relation between emissions and
both time t and turbulence as characterized by (3.

     The effect that micromixing of this nature has on emissions is shown schematically in
Figure 7. In the first  diagram, the fluctuation in fuel concentration  with time is shown.
Because concentration fluctuations will eventually decay (according to the Corrsin model),
then as long as the mean concentration C falls between Cs and Cq, the entire distribution of
fluctuations will  eventually fall  into  the combustion  zone and the emissions will  be
eliminated. To promote  this change more rapidly, one can follow two different courses of
action:

     •   Shift C relation to the constraining boundaries Cs and Cq, or

     •   Increase 0t to achieve a faster rate of decay in the fluctuations.

     The first approach  to  eliminating  emissions on the microscale is shown in the second
diagram of Figure 7.  Whenever the mean concentration is close to Cs (or alternatively to
Cq), then the emissions will tend to be high. However, by adjusting the mean concentration,
a greater proportion  of the fluctuations will occur within the  combustion zone and the
emissions will be diminished. Since the concentration is directly related to the temperature,
the importance of maintaining the temperature within the combustion range is clearly seen.
This is the primary effect  of  temperature on  emissions.  The second approach focuses on
achieving more complete decay of the concentration fluctuations.

     In  the third diagram,  a  typical concentration fluctuation for  a relatively unmixed
system would appear  as the solid lines  and would result in a given amount of emissions at a
specified concentration  (or temperature) and |3t. By maintaining the mixture at the same
concentration, but improving the level  of mixing, the fluctuations can be reduced as shown
by  the  dotted profile  so  that  the emissions  can be reduced or eliminated.  The  same
argument also applies to increases in residence  time at constant  temperature and degree of
turbulence.  Although the singular importance  of adjusting stoichiometry is crucial to the
reduction of combustible emissions, these secondary relationships between temperature and
both turbulence  and time have also been noted in order to  complete the quantitative
description of the interrelationships among the three T's.

     Application  of Micromixing Models to Incineration —  The previous two headings of
this section dealt with the  development of a micromixing model to correlate emissions data
                                         29

-------
OJ
o
                         ^STOICHIOMETRIC

                           .EMISSIONS
       MEAN
Q
                                     ENCH

                                  EMISSIONS
                         NATURAL DECAY
                                                       ADJUSTMENT OF STOICHIOMETRY
PROMOTION OF DECAY
                                                       Figure 7   Overall effect of micromixing on emissions.

-------
with furnace operating parameters, particularly the three T's. The two parts of the model
that were presented are:

    •   Use of the unmixedness factor, >/c'2  to relate emissions data to fluid
         mechanics properties of the system; and

    •   Use of Corrsin's Equation to relate the fluid mechanics properties to time,
         temperature, and turbulence.

In this section, we have outlined the  steps required  to  apply both parts of the model to
incinerators.

    Throughout this part  of the report  it is necessary to understand the  combustion
phenomena at work within the incinerator. In the previous mixing study (2), we developed a
model to characterize the bed burning process. During the course of this work, it became
more  and more apparent that the characterization of bed combustion was not one of the
primary factors necessary  to  understand and control  combustible emissions. However, the
bed model was often necessary to estimate the value of the parameters used in the mixing
models.  The  development of  the bed burning model is given in Appendix E. Some of the
results of this development are summarized below.

    Within  the refuse bed, temperatures are high enough to assure  complete burning of
pyrolysis or gasification products that come into intimate contact with oxygen. More likely,
however, there will be insufficient oxygen or poor mixing so that combustible gases will be
given  off.

    The offgas contains two  distinct constituents: (1) combustibles and combustion prod-
ucts,  and (2) air. The composition  of the combustibles and  combustion products  is
controlled by the water gas shift equilibria. The composition can be estimated (as shown in
Appendix E) if the initial value  of CO/CO2  is known. Typical compositions are shown in
Table 1 for different values of CO/CO2. These are also shown graphically in Figure 8.

    The actual value of CO/CO2  was determined empirically to be 0.21 (range 0.02 to 0.6),
based on  test data taken  directly over the bed of an existing incinerator. The data are
presented in Section VII.

    The offgas from the refuse bed also contains oxygen that has not yet been consumed.
Since oxygen and combustibles do not co-exist in a well mixed state, the oxygen available in
the offgas must be contained  within a  pocket of unmixed air. The  fraction of such air in the
offgas is simply the ratio of available oxygen in the sample (in percent) to 21%. In the tests
reported in Section VII, the incinerator contained between 29% and 90% unmixed air with
an average of about 50%.
                                         31

-------
                         Table 1. COMPOSITION OF BED OFFGAS
CO/CO,
 0.25
0.50
 1.0
 1.5
 2.0
Analysis (vol %)
C02
CO
H2O
H2
N2
13.5
 3.4
22.4
 2.8
57.9
12.2
 6.1
22.0
 5.5
54.2
10.5
10.5
21.0
10.5
47.5
 9.3
14.0
19.9
15.0
41.8
 8.4
16.8
18.9
18.9
37.0
02  (req.)              3.1
mol 02/100 mol offgas
                5.8
               10.5
             14.5
             17.8
                30
                20
             LLJ
             O
             oc
             111
             Q.
                10
                               0.5
                                                I
                          1.0
                       [CO/C02]
                          1.5
                        2.0
                             Figure 8   Offgas composition of CO/CC-2
                                            32

-------
     • Application of the Unmixedness Factor

     In Appendix B, the procedure for calculating the unmixedness factor from a stack gas
analysis is given. It is also possible, given an unmixedness factor, to calculate the emission
factor if the initial CO/CO2 ratio is known. In Figures 9, 10, and 11 are shown three sets of
calculations based upon different CO/CO2  values and also different empirical relations for
CO2 vs temperature. In this work both the initial CO/C02 ratio and_the stack gas analysis
were  measured. Hence it was possible to determine the value for >/c'2 two different ways
and compare the resu'ts. This is also done  in Section VII. The good agreement between the
experimental value of >/c'2 (based on gas analysis) and the theoretical value of\/c'2 (based
upon the CO/CO2 ratio) supports the contention that emissions can be determined using the
unmixedness factor approach.

     • Application of Corrsin's Equation

     To apply Eq. (21) above to incinerators, one must be able to estimate the following
parameters:
The first two are related to the aerodynamics of the furnace mixing and also to the furnace
residence time (for gases, not refuse). The latter two parameters require the knowledge of
the composition of offgas emanating from the refuse bed. The estimation of each of these is
discussed below.

    Parameter 0 — For the incinerator, the  only types of mixing energy that have been
considered  are  those introduced  by incoming air  streams.  Equation (22) can then  be
expanded as follows. The rate  of input of mixing energy for a given air stream is calculated
from the equation:
              d (KE)     u mil2               „        ~*,  ,    <
              u_£_;  =  AJ321L  =  '/2(pAu)u2  = V2pQ3 A2  •                 (23)
               dt         At

If each stream is evaluated at the temperature  at which.it gives up its energy to create
turbulence, then as a conservative estimate each jet must be evaluated at the jet tempera-
ture,  while the gases coming off the refuse bed are evaluated at the furnace temperature.
The energy term for a jet acting over a portion of the grate is:
          e =                     p
                                    ^                     :

                                        33

-------
      2500
 CM

8
o
o


Q.
tr
O

o
to
CO
      2000  -
      1500
      1000  -
       500 -
         750
1000
                                                                        1750
                       1250            1500



                         TEMPERATURE, °F



Figure 9  Dependence of emission factor on temperature and unmixedness factor.
                                                                2000
                                                34

-------
      3000"
      2500 -
 CM
O
CJ
S9
CM

@
O
O

a
a
ir
O
u
2000 -
—    1500 -
      1000 -
           BASIS:
             C0/C02 = 0.5
             %C02 (WET) =
               
-------
     2000
     1500  ~
 CM
8
8
a
cc.
O
z
O
1000 -
      500  ~
             BASIS:
             C0/C02 = 0.50
             %co2|w
                    = (TEMP-750UF)/1250UF
        1000
                     1250              1500

                               TEMPERATURE, °F
                                                             1750
2000
             Figure 11  Dependence of emission factor on temperature and unmixedness factor.
                                            36

-------
By  substituting in the  ideal gas law and  scaling all dimensions with  the  characteristic
dimension L, which is taken to be the shortest dimension of the furnace (usually the furnace
width), the equation for j3 becomes:
                                                       QJ
                                                            Qo
                                                                       1/3
                                                                                (25)
where
         7    :

     subscripts
              Lf/L   =   ratio of length to width,
              Hf/L   =   ratio of height to width, and
         o    =    initial condition at top of bed
         j    =    Jet.

If uniform mixing has not been achieved, then j3 will have a different value for each furnace
streamline.

     Parameter t  — Parameter t is the  residence time of combustion  gases within the
furnace. One method for calculating t was proposed in the section on fluid flow [see Eqs.
(11) through (13)]. Another approach is to relate t to the gas flow rate Qf and the furnace
volume Vf.  This  could be  combined with Eq. (25) above to calculate j3t according  to the
equation:
                       = 0.15
assuming that
                                                                           (26)
                                      t =
Note that without jets (0t)0 equals 0. 1 5 y2 / 3 .

In this case, the dependence on furnace length is eliminated. Since macroscale uniformity is
more probable across the width of the furnace or from grate to roof, but not from back to
front, the use of this form of the equation is more consistent with the local character of the
microscale model.
Parameters
                          and AC — The effect of overfire air jets can be determined from
a material balance of the jet mixing process using the calculated  offgas composition for
                                         37

-------
refuse beds (See Appendix E). Consider for example how the composition of bed offgas is
affected by dilution with high-velocity air jets. The following equations are based upon the
point of jet mixing as sketched below:
                           Jets
                                ..;;..- Diluted-.;
                                "'.'.'•'.. Gases ".
                                           •*>-
                                              Jet Mixing Zone
                                 •.."..•Offgas .-.
                                   Refuse Bed

Let x equal the mole fraction of offgas in the diluted gases. Then x can be estimated in one
of two ways. The amount of carbon in the offgas is the same as the amount of carbon in the
diluted gases. At the breech of the incinerator essentially all of the carbon is in the form of
CO2.  If a gas analysis is  taken before and after the jets are turned on, and  if subscript b
represents the base condition (before jets) and subscript j represents the jet condition, then
the carbon concentration represented by CO2 is:
                                     lj = x(%C02lb)
or
                                    x =
                                       %C02lj
                                       %co
                                             (27)
                                             (28)
                                            2'b
Alternatively, a balance of total moles at the point of mixing yields the equation:

                                PQi       PQC      PQf
                                         RT,
               RTf
                                                                                 (29)
In this equation, the variables that  can be measured are Qj, Tj, Qf, and Tf. The defining
equation for x then becomes:
x = 1 -•
                                           Qf/Tf
                                                                                 (30)
     The  factors required to calculate yc'2 all change  with the  addition of jet air. The
changes in these variables are calculated using the following equation. The change of )3t is
best determined based upon Eq. (30):
                                  1 +
                                        (Af/Tf)
                 0-x)3
                                                            1/3
(31)
                                         38

-------
Addition of air will increase available oxygen according to the equation:
                           (%02 )j = x (%02 )b + (1-x) 21%  •
                                                                            (32)
Likewise  the oxygen  required for complete combustion after dilution is determined from
the equation:
                           (O2Req)Jo =(02 Req)bo-x
                                                                            (33)
Dividing Eq. (32) by Eq. (33) and combining terms, one obtains the following:

                                                  l-x\  21%
                /   %0a\     / %02  \

                \°2Req/J0    \02Req/b
                                             1 +
                                                       (%02)b
                                                                            (34)
Note that Eq. (34) applies only to the region of jet mixing (subscript o), not to the entire
streamline. A similar approach is taken to determine Ay. In Appendix B it is shown that:
                                          %CO2
                                                                             (35)

                                          %C02lb\
                                          %co2ij
where y represents the mole fraction of original combustibles (before air entrainment within
the bed). The addition of jet air carries the dilution of the offgases even further than occurs
within the bed itself. At the exit of the bed:
                                              %C02b
                                                                             (36)
After the jets:
                             (y-ys)j = ys
                                            /%CO2I

                                            \%CO2
The equation for Ayj can be expressed in terms of x as:
                                            '%C02 b

                                            %C02
                                                     (1-x)
                                                                             (37)
                                                                             (38)
The above equations can be combined with Corrsin's Equation and Figure 3 to provide a
single model equation as follows:
From Figure 3:
                                   '2
                                   Ay
                                                                             (39)
                                       39

-------
 Note that f (00) is given in Appendix F.

 Combining this with Eqs. (34) and (38) yields:
     The jet effect can be graphically illustrated in the  following way. Figure 10  can be
 replotted (Figure 12) to  show emission factor  vs \/c'2  (rather than temperature). Then
 Point A  on  Figure 12 represents some base condition as determined in the breeching of the
 furnace.  The initial base condition can be determined by  applying Corrsin's equation in the
 form:
 The initial  point  can  be located  on Figure 12 by assuming  that the initial and breech
 conditions are at  the  same temperature, i.e., the decay process is isothermal as shown in
 Figure 1 2 (Point B). One could also apply the model by specifying the initial temperature
 rather than making the isothermal assumption. The initial jet condition can be determined
 by noting that the emission factor does not change with dilution, but remains constant in
 going from the base to the jet condition. Because of the isothermal assumption, the factor
(v/c2 Y   is  the  value  determined  at the emission factor  of the base condition and  the
V   /Jo                                                 /  r=.\
 temperature of  the jet condition, i.e., Point C. The decay of(-^/c  )j  is also isothermal, so
 that the final value will  lay somewhere along the jet isotherm, Point D, for example. The
 magnitude of decay can be determined from the equation:
                                                                                (42)
This equation is the basis for predicting the effect of jets on combustible emissions.

Estimated Effect of Jets on Incinerator Emissions

     The effect that jet mixing has on combustible emissions follows from Eq. (40) above as
seen in the following example.

     Consider a furnace initially at 1800°F, having an unmixedness factor of 0.20 and with
j3t equal to 0.25. The relationship between emissions, temperature, and y^c5 is assumed to
be that of Figure 9. The empirically derived relationship between temperature and %CO2  is:
                          %CO2     x(%CO2 b)    T-960°R
                          	 =  	  =	
                           10%        10%         1500°R
                                        40

-------
     3000
     2500
     2000
CC
O
O
     1500
     1000
      500
                                                                                                                    INITIAL VALUE
                                                                                                                       WITH JETS
                                                                                                                 EFFECTOR JETS
                                                                                                                 (SEE TABLE 2)
                                                                                                    FINAL VALUE
                                                                                                     WITH JETS
                          .05
                                            .10
                                                             .15
.20
                                                                                                 .25
                                                                                                                   .30
                                                                                                                                     .35
                                                                                                                                                       .40
                                   Figure 12  Dependence of emission factor on temperature and unmixedness factor (revised plot of Figure 10)

-------
At 1800°F (2260°R), %CO2 is 8.7%. The ratio of Af|Aj is taken to be 150. A summary of
the calculated terms for the jet  conditions is given in Table 2 based upon the following
calculations:

              Parameter                       Source

                                              Given

                                              Eq. (27)

                                              Eq. (43)
                                              Figure 3
                                              Eq.(31)

                                              Eq. (40)orEq. (41)

             L/d* L - U/c2   }              By substitution of Eq. (42)

Note that the internal consistency of the equations can be demonstrated by calculating the
final two parameters listed above, using either of two equations each.
     The values for ye'2 vs temperature given in Table 2 have been plotted on Figure 12.
The jets have a dramatic impact on reduction of emissions whenever x is less than 0.95. In
Section VII, we have shown that other conditions could be set forth in which the jets prove
to be completely ineffective or of marginal value.
                                        42

-------
Table 2.  SUMMARY OF CALCULATED TERMS


X
Base
Jet







X
1.0
0.98
0.95
0.90
0.85
0.80
0.75


T,°F
1800
1778
1740
1675
1609
1544
1479


*o
27
28
30
34
38
43
48

/ /Tf\
(v£E)
\ AC/o
0.774
0.765
0.750
0.725
0.700
0.680
0.665


AYO
0.332
0.341
0.355
0.377
0.400
0.422
0.445
/ \
/ /=j?\
r>
0.257
0.261
0.266
0.273
0.280
0.287
0.296
C"~ \ •,
150 .T:\2
1 1 ,
Tf ) <1-*> J
1.0
1.004
1.055
1.356
1.853
2.442
3.032
1/3 ^_
/&
Vc
0.200
0.203
0.204
0.195
0.176
0.156
0.137

A»/c^ *
t* V
0.057
0.058
0.062
0.078
0.104
0.131
0.159

-------
                                    SECTION VI

                            EXPERIMENTAL PROGRAM

     A large portion of this program was devoted to testing an existing municipal inciner-
ator to:

     1.   verify the validity of applying the basic combustion fundamentals presented
         in the previous chapters to the design of municipal incinerators;

     2.   demonstrate the effectiveness of overfire jets in reducing combustible emis-
         sions; and

     3.   provide  experimental data on municipal incinerator operation in hopes of
         promoting further technological advances in design or operation.

The test equipment procedures and data reduction methods are discussed in detail below,
with the data tabulated in Appendix A.

TEST INCINERATOR

     The primary piece of equipment used in the test was a municipal incinerator. Before
selecting the test vehicle, we reviewed the various types of incinerators in use in the United
States today. We limited our  selection  to grate-type incinerators, because the batch-type
incinerators are being phased out around most of  the country and suspension-fired units are
not in sufficient use to have general applicability. We considered several local  incinerators
before finally selecting the incinerator of the City of Newton, Massachusetts.

     The Newton  incinerator  is a 500-TPD capacity  plant  consisting of two  identical
250-TPD traveling grate furnaces as shown  in  Figure 13. The incinerator configuration is
typical of many found throughout the country today. The refuse arrives at the facility and
is  weighed before being dumped into  the  loading pit.  A  fine  water mist  is  periodically
sprayed over the refuse to control the dust. Unfortunately, much of this water settles to the
bottom  of  the pit so that the refuse there is very moist and burns poorly. Both  furnaces are
fed by a 3.5-ton traveling bridge crane provided with a 2-yard clamshell bucket. The charge
chute to each furnace is water-cooled and gravity-fed. During operation, the crane operator
is  reasonably careful in blending the various types of refuse to obtain a uniform mixture.
Even so, large fluctuations in the refuse composition are seen  so that consistent operation
on a particular type of refuse is  difficult to maintain for more than one or two hours. In
collecting data on emissions, we  preferred test runs lasting for as long as 4 hours, but in
many cases changes in refuse composition or other adverse conditions forced a change in the
operating conditions of the incinerator.
                                         45

-------
        ROOF £L. 92.0'
        DUMPING AMD STOKII6
         FLOOD EL. 63.0'
        GRADE EL. 60.0*
O\
                  "H
                                                 Figure 13  Cross section of Newton 500-TPD Incinerator showing sampling locations.

-------
Furnace

     Each of the rectangular furnaces is lined with conventional refractory, except for the
lower portion of the wall about three feet above each grate which is lined with a silicon
carbide brick to resist abrasion and slagging. The distance from the grate to the roof at the
charging end is approximately 10 feet; at the drop-off end it is approximately 30 feet.
     The furnace has three traveling grates to convey the refuse through the incinerator. The
initial grate is approximately 11 feet long and is inclined to facilitate radiative transfer of
heat to dry the incoming refuse.  Hence, it is designated the drying grate. The refuse is drawn
from the feed chute into  the furnace through an opening 4.2 feet high by 8 feet wide. The
feed rate of refuse into the  furnace can, therefore, be determined by the speed of this first
grate and the density of the refuse. For an average density of about 15 lb/ft3, a grate speed
of 40 ft/hr results in a feed  rate of 250 tons/24 hour  day, which is the design capacity. The
grate itself consists of rows of overlapping keys through which air is passed both to promote
burning and to cool the burning surface. The keys  are made of cast iron  to resist high
temperatures.  The rack bars holding the keys are made of mild  steel. Throughout the
program  conditions often existed that, if not  properly adjusted,  resulted in overheating of
the steel portions of the grate. Since  a severe  overheating could have caused damage to the
grates, these conditions were abandoned whenever  the  grate temperature  began to rise
abnormally.

     At the end of the drying grate, the refuse tumbles on to the second or burning grate. The
vertical distance between the  two  grates is approximately 5 feet which is sufficient to promote
a certain amount of refuse mixing, which tends to expose additional unburned refuse to the
flame. The second  grate is 16 feet long and slightly  inclined.  Its speed can also be
controlled as can the speed  of the third  grate. In general, the second grate is operated at a
somewhat faster speed than the  drying grate because  past experience has shown that when
the grate speed is equal or  slower than the drying grate, the latter tends to overheat. The
most probable  explanation for this  phenomenon is that the slower grate speed of the lower
grate results in a buildup of refuse directly under the nose of  the upper (drying) grate.
Then, either because of a  restriction in the air leaking into the furnace which normally cools
the nose of the grate, or because of increased heat generation at that point on the grate, the
grate overheats. Since this portion of the grate is not cooled by underfire air, the grate will
return to the  charge chute  hotter  than  normal. When this occurs, the refuse on the grate
eventually  reaches a point at which  it could  be prematurely ignited. With the flame right
next to the grate surface, an abnormally high heating effect occurs which is not relieved by
the underfire air. This  phenomenon  happened several times during the course  of the tests
and prohibited  operation for long periods of time at relatively slow burning grate speeds.
                                          47

-------
     The refuse  is tumbled for a second  time from  the  burning grate to  the  third or
 burnout  grate. The flame on  this  grate  is  characteristically  lower in  intensity than
 on the second grate and,  in many  cases, the flame has  either  been extinguished or is
 merely smoldering by the time the refuse has traveled the 16-foot length. One characteristic
 of this phase of the operation is that many of the slower burning items, such as heavy books
 or logs, are still burning when they reach the end of the grate. To provide a longer residence
 time for the furnace, this grate was typically operated at speeds equal to or slower than the
 second grate.  Unfortunately,  at  slower grate  speeds the residue tended  to build up,
 restricting the air flow through the bed and, consequently, reducing the combustion inten-
 sity. A consistent problem throughout the test was the development  of methods for
 operating this grate to achieve good burnouts, while at the same time  maintaining temper-
 atures  above it that would  be well above the quench temperature. The latter is extremely
 difficult  and   consequently temperatures over this  grate  as low  as  800°F were  not
 uncommon.

     When the refuse reaches the end of the burnout grate, its ashes fall into a water-filled
 hopper where they are quenched.  The residue is next transferred to a  truck by means of a
 continuous-chain   flight conveyor, and eventually is used as  landfill.  The overall volume
 reduction is typically between 80% and 90%.

 Combustion Air System

     The  primary source  of combustion air is a forced draft fan rated at 40,000 cfm (6
 inches SP) and fitted with an inlet vane control. The underfire air  is supplied to. the grates
-by three  separate ducts running from  the main fan discharge to the air plenum located
 between the upper (forward) and lower (return) sections of each of the grates. Each plenum
 is divided into two  sections, with the distribution controlled  by  a manual damper at the
 entrance  to each system. Air  is  forced through the  upper portion of the grate,  passing
 through the keys and cooling the burning surface. The boxes are not particularly well sealed,
 so that high static  pressures in the wind  boxes cannot be attained.  For this reason the
 amount of air that can be forced through the refuse bed is severely  constrained, particularly
 in the  drying grate. The third grate has very low refuse bed depths  at normal or high speeds
 and can accept relatively large portions of air without difficulty. This configuration often
 results in  high excess air over this grate and consequently low temperatures.    ;

     The  overfire  air is  taken from the fan discharge using only one large duct with a single
 manual damper. The overfire air is introduced into the furnace through three banks of roof
jets consisting  of  five  6-inch diameter jets each. The  primary purpose of this overfire air
 system is to cool the overfire region. The overfire mixing system,  installed for emission re-
 ductions, had a greater number of jets operating at higher jet velocities.

     Additional air is used to cool  the silicon carbide  portion of the wall to prevent slagging.
 The air enters the furnace  at approximately 3 feet above the grate level.  The forced draft fan
 supplying this air is rated at 4500 cfm (6 inches SP). This source amounts to less than 10%
 of the  total air flux into the  furnace.
                                          48

-------
     By far the largest source of air is in the form of uncontrolled leakage into the furnace.
The leakage rate was calculated by material balance and, typically, ran as high as 30,000 cfm.
A  large fraction of this air is drawn in  from beneath the grates and is a natural source of
cooling for both the grate and the overfire region.

Stack Gas Removal

     Each furnace is equipped with a fly-ash removal chamber containing water sprays and
fire  brick baffles.  The hot  combustion gases  pass through  the  breeching area of the
incinerator into these chambers. The gases are cooled to about 350°F and the large fly ash is
removed.  Water from  the baffles  is collected in a sump, clarified, and recirculated. The
baffle  chambers merge directly into a single natural draft chimney which is 175 feet high
and  designed  for the  gas  flow  required to accommodate the ultimate capacity of both
furnaces. Because the cooling of the baffle system maintains the exit temperatures in a fairly
narrow temperature range of approximately 250° to 450°, the chimney draft that results is
relatively constant.  For this reason, both the draft pressure in the furnace and the gas flow
rate  through  the  furnace  remain constant. Hence, the air distribution could be  somewhat
controlled, but not the  flow rate. This constraint was a constant source of operational
difficulty throughout the tests.

Incinerator Control

     To control the incinerator  at the  proper conditions, additional instrumentation was
required to augment its present monitoring system. The grate speeds were  determined with
the existing rate meters, and the speeds  were verified to ensure that the meters were reading
properly.  Pitot tubes  were installed in each of  the  underfire air ducts and in the main
overfire duct to record the air input into the furnace. The total air volume was checked with
a pitot tube reading at the inlet  duct  to the main fan. In addition, the leakage air was
estimated using a hot-wire anemometer  to determine the air velocities at the most probable
points of entry.

INCINERATOR MODIFICATIONS

     To follow the strategy for eliminating combustible emissions on both the microscale
and the macroscale, one of the furnaces was modified with an overfire jet system. Based on
information gathered during the initial test of this program, it was clear that the gases from
grates  1 and 2 were very hot and oxygen-deficient, while the gases coming from grate 3 were
at or below the quench temperature. To demonstrate the design principles and effectiveness
for jet systems,  we installed  sidewall jets over  grate 2 to  satisfy  the  stoichiometry suf-
ficiently and also to achieve sufficient penetration. The forced-draft fan that was selected
had a design  capability of 15,000 cfm at 6 inches SP. To achieve the greatest flexibility and
control, the fan was equipped with both inlet vane control and an outlet damper.
                                          49

-------
     The  duct network consisted of  16-gage welded tubing and channeled air from the
outside through the fan to a Y-connection, where the air was split into two streams; each
stream was delivered to a jet manifold located on opposite sides  of the furnace. The air
distribution was monitored using a pitot tube in the main inlet duct and pitot tubes in each
of the two ducts feeding the jet manifolds. In  addition  to the  controls on the fan, a
butterfly damper was installed in each of the two manifold ducts.

     The manifold system consisted  of a long header off which 10 duct sleeves ran, carrying
air to the 10 jets on each side of the furnace. Each tubular sleeve had a 5-inch diameter and
was equipped with both a 1/8-inch velocity pitot tube and a butterfly damper for velocity
control and monitoring. Because of the experimental nature of the program, the duct sleeves
were connected to the jets by flex-hose connections. During the course of the tests, the flex
tubing caused many maintenance problems. Thus,  this type of construction is not recom-
mended for permanent installations.

     The jet nozzles were constructed of  5-inch,  schedule 40,  mild steel pipe, inserted
through the wall of the furnace to within 2 inches of the inner refractory wall. The external
portion of the pipe was connected to the duct sleeves by means of a flanged connection. In
installing the jets, the location for each jet was marked on the inside of the furnace, and the
hole was made using standard techniques. The jet pipe was inserted through the wall, welded
to the outer furnace shell, and surrounded with a high-temperature insulation. The inner
refractory wall was then molded  around the jet using a plastic refractory. During the course
of the tests, slag buildup around the  jets  was quite common. Therefore, a permanent jet
installation  should use precut silicon  carbide refractory or other slag-resistant material to
avoid this problem.

CONDUCT OF TESTS

     The composition of refuse that is burned in the average municipal incinerator varies
from  hour  to hour  and day to day, making  it  difficult to arrive at any  consistently
reproducible test conditions. During the operation  of the Newton incinerator, these varia-
tions could  be seen in breeching temperatures fluctuating as much as 500°F within a period
of 30 minutes; variation in the breeching temperature of ± 100°F was considered a relatively
stable condition. It  was therefore difficult to measure precisely all of the variables affecting
the system.

Test Parameters

     Test  variables included  the refuse feed rate, moisture content,  the underfire and over-
fire  flow  rate and  distribution,  and  the refuse residence  time. Unfortunately, the  test
variables are interrelated and  are often dictated by the permissible or possible conditions
under which the incinerator can be operated. Each of these parameters is discussed in de-
tail below.
                                         50

-------
Feed Rate

     The refuse feed rate is controlled by the grate speed of the drying grate and it typically
runs between  35 ft/hr and 55 ft/hr, being set according to whether the refuse is relatively
wet or dry. The volumetric rate in cubic ft/hr is a product of the grate speed in ft/hr and the
area of the charging port (33.6 ft2).

     The weight of refuse depends upon  the refuse  density which often varies. The density
of Newton's refuse was reported previously by Kaiser's[8] to be 13 lb/ft3, but similar data,
both at Newton and for other incinerators, indicate densities  as high as 17 lb/ft3. In the
calculation of feed rates there is an additional uncertainty due to the packing of the feed
hopper; if it is packed loosely with large voids, the apparent bulk density will be lower than
13 lb/ft3. On the other hand, a tight packing will result in a higher apparent bulk density.
The density also varies with moisture content, tending toward the higher values for wet
refuse.

     In the conditions which were used in the current work, the following feed rates were
utilized:
Grate speed, [ft/hr]
Volume feed rate, [ft3/hr]
Mass feed rate,* [tons/hr]
Heat release,** [MM Btu's/hr]
35
1176
7.7
68-78
40
1344
8.7
78-87
45
1512
9.8
88-98
  50        55
1680      1848
 10.8      11.9
98-107    108-118
  *P= 13 lb/ft3
**heating value of refuse = 4500-5000 Btu/lb,  100% burnout.

Moisture Content

     Refuse will always contain  moisture but it  can at times have considerably more
moisture than usual. Typical values range between 20% and 50%. At Newton, the higher
moisture conditions result usually  after a rain, or when the refuse is taken from the bottom
of the pit.  To control the dust,  the pit is periodically sprayed with water while trucks
unload. Excessive moisture drains to the bottom of the pit. At times water was observed
dripping from the clamshell bucket as refuse from  the bottom was charged. The effect of
moisture was  clearly shown during the runs with the amount of refuse moisture changing
over the course of a day's run as the crane worked from the top to the bottom of the pit.
                                         51

-------
     In describing refuse, we have qualitatively used the labels of "wet" or "dry," but we
recognize that the absolute moisture content is not such  a discrete function.  We  also
recognize that our visual observations could be quite misleading, and that the data on the
combustion products could contradict our initial moisture labels. Of the principal variables,
this one is by far the most uncertain and least controllable of the operating variables.

     The moisture content was allowed to vary normally, resulting in conditions for the first
tests  fairly  evenly  divided between "dry"  and  "wet" refuse.  Some of the runs reflect a
change  from dry to wet during the progress of a test from beginning to end. The  final series
of tests demonstrating the use of overfire jets had primarily dry refuse.

     Unfortunately, the refuse moisture content, which drastically affects the refuse heat
value, also affects very much the allowable feed rate. Wet refuse is normally burned  at a
slow refuse feed rate to achieve an acceptable burnout. If higher feed  rates are attempted,
the ash level on the burnout grate tends to build up, and the ash conveyor system becomes
overburdened.  Continuous operation in  this mode during some of the tests resulted in
temporary shutdowns because of the strain on the ash conveyor system. On the other hand,
by reducing the second and third grate speeds a good burnout could be attained, but then the
furnace would overheat. Hence, wet refuse  was almost always run at low grate speeds.  Dry
refuse, on the other hand, was normally run at high grate speeds, and we found that at low
grate speeds most of the combustion was  completed on the first and second grates, with
little  combustion taking place on the burnout grate. However, it was still possible  to vary
feed rates for dry refuse between 35  ft/hr and 55 ft/hr with only minor interdependences.

Air Flow Rates and Distributions

     Because of the natural  tendency toward a constant furnace air flow induced by the
natural-draft stack, the effects of varying air flows and distributions were difficult to detect.
However,  various air distributions are used primarily in the underfire air  system  where the
air flow rate is related to burning rates. Since most of the leakage is into the overfire region,
these effects were relatively unaffected.

     The  underfire air  can be divided into six discrete streams, since each underfire  air
plenum has two sections. In practice, controlling dampers become clogged with ash reducing
the effectiveness of control.  Care was taken during tests to ensure  that these dampers did
operate effectively, but some reduction of control was still observed.

     The quantity of air flowing under each grate was measured by means of a pitot tube.
The flow  was initially calibrated by complete traverses of each duct, and periodic checks
were  made  to  verify the continued accuracy of the sample point readings. While it was
hoped that  the underfire air would be in excess of stoichiometric  air, this  could  not be
readily achieved, particularly under the first two grates.
                                         52

-------
     The distribution of  air was  not as controllable as desired either. Only very small
amounts of air (5% stoichiometric) could be forced under the drying grate without causing
the draft to drop to zero (or slightly positive  pressure), forcing smoke and fire out of the
furnace. The second grate underfire air generally accounted for 15 to 30% of the stoichio-
metric requirement, while the remaining 35-40% of stoichiometric air fed under the grate
came through the  third or burnout grate.  Because of a combination of buoyancy effects
(larger drafts  at lower  elevations  and a lower pressure drop through the residue of the
burnout grate), it was difficult to reduce  the air under that grate significantly. In addition, a
large amount of leakage was present around this end of the furnace.

     The two  existing sources of controlled overfire  air are the roof jets, used primarily for
furnace cooling, and the wall-cooling air. During most tests the roof jets  were off, and the
rate of the wall-cooling air was consistently about 4,000 cfm.

     The overfire jet system described earlier could supply up to an additional 20,000 cfm.
It is interesting to  note  that at high jet flow rates the leakage was noticeably curtailed. The
various types of flow distributions common to the jet system are discussed in detail in the
section covering test results (see Section VII).

     The remaining flow was uncontrolled resulting from  the natural  tendency  of the
furnace to  draw air from  the outside. The leakage was great typically between 20,000 and
30,000 cfm, but under the right  conditions this could be reduced to a very low flow rate. In
fact, very high forced air  flow rates would result in almost zero leakage, but would effect a
noticeable increase  in the amount  of smoke emissions onto the operating floor. In addition,
the reduction  in leakage usually  resulted  in an increase in the nose temperature of the grates
(loss of natural cooling). These factors effectively limited the amount of underfire air that
could be used.

Furnace Residence Time

     Each of the three grates can be controlled independently so that the residence time for
the refuse is simply the sum of  the residence times on each of the three grates. Under most
runs the residence time  was approximately one hour. Certain combinations of grate speeds,
however, are not acceptable, as  discussed previously,  because of the tendency of the drying
grate to overheat. For most runs the three grate speeds were set at the same speed, but some
tests  were  made  at relative residence  times both  slower  and faster  than the  normal
condition.

Test Sequence

     In each test the run conditions were  first established and  the incinerator was allowed to
reach steady-state operation before test data were taken. During the first 30 minutes to one
hour, the various parameters were monitored and final checks were made on the instrument-
ation to ensure that no major difficulties would be encountered during the run. This waiting
                                          53

-------
period  was necessary to allow  the  furnace  to  reach  steady-state  operation  at  the  test
conditions,  and was approximately equal to residence time of refuse in  the  incinerator.
Although two or three residence times might be desirable from a theoretical standpoint, the
large variation in the refuse  composition over the course of a run was enough  to  obscure
minor deviations from the true steady state and additional waiting time was not warranted.
In the initial sets of tests used for verifying the analytical methods,  data were taken over a
period  of 4 to  5  hours.  Because of improvements in the instrumentation and analytical
procedures, most of the data for the jet tests were taken over a period of one-half hour on
most runs. In this time two complete sweeps of the three breeching zones could be achieved,
corresponding to two sequential  15-minute average moisture determinations. The gas analy-
sis was recorded before and after the test and trends were  compared to assure  that the gas
averages were reasonably representative of the overall operation. The results of these tests
are presented in Section VII.

Input Data

     During the course of each test the following data were recorded:

     o    The furnace control panel readings,  which  include the  furnace breech
          temperature (high  temperature),  the stack temperature, the furnace draft
          pressure, the chimney draft  pressure, the forced-draft fan pressure,  the
          under-grate pressure, and the grate speed for each  of the three grates;

     o    The air distribution and flow rate, including velocity measurements for each
          of the three underfire ducts and the overfire duct;

     o    The velocity of each jet, as well as the total flow to the fan and to each jet
          manifold;

     o    The static pressure  underneath each grate (during selected runs), and

     o    Overall furnace  conditions and specific events (taken chronologically) which
          might  affect the test data or furnace performance, including any shutdowns,
          changes  in refuse feedstock or furnace operating conditions, the degree of
          burnout, and the approximate condition of the refuse feed.

     The emissions data and the data reduction techniques used are discussed below.

Emissions Data

     Measurements of emission  composition, temperature, and flow were  made in  the
breech and over the bed. Further measurements of static pressures were made at selected
points to  allow estimation of air leakage into the incinerator. Particulate emissions were
made in the  stack  during the initial  baseline  series of tests and are reported in detail in
Section VII.
                                          54

-------
     Output flows were calculated from  the breeching area velocity data collected  using
S-type pitot tubes on the gas-sampling probes and the temperature measurements.  Com-
position data were obtained to measure  the overall  combustion characteristics and the
combustible emissions. The species selected for measurements were:

       CO2 - carbon dioxide               H2  - hydrogen
       O2   — oxygen                      CO  — carbon monoxide
       N2   — nitrogen                     HC  — total volatile and gaseous hydrocarbons
       H2 O - water

     In addition, during the particulate measurement samples were collected for

       Cl   — chloride (as hydrogen chloride),
       SOX - total sulfur oxides, and
       NOX — total nitrogen oxides.

     The collected particulate was also analyzed for trace metals.

     Measurements were made at several points in the incinerator system. A cross section of
the incinerator is  shown in Figure 13. The majority of the measurements were made in the
breeching at points D, E, and  F. During studies of the bed-burning model, composition
measurements were obtained along the bed at points A, B, and C. The continuous-stack CO
monitor drew samples from point H. Initially velocity measurements  were made only at the
points of  gas composition measurement,  but in the final jet-mixing phase of the program,
velocity and pressure measurements were made at points A-H as discussed later.

     Initial  experimental  design called  for operating  the  incinerator  for  an  entire day
(5-6 hours) at a  single condition to normalize  refuse-variation  effects. On that basis the
analysis system was designed on a  time-sharing basis so that  a single set of analytical
instruments could be used to measure the composition  of three  points in  the breeching or
furnace. Velocity  and pressure measurements were also time-shared on a single instrument,
while temperatures were all recorded simultaneously on  a multipoint recorder. Problems
with obtaining accurate moisture analyses necessitated the use of an  independent sampling
line piggybacked on the gas probe for H2O determinations. The general overall measurement
approach is represented schematically in Figure 14.

Analytical Procedures

     The basic sampling systems are shown schematically in Figure  15. There  were  three
such identical systems connected to the gas analysis instruments on a  time-shared basis. The
probes, cyclones,  and H2 O silica traps were mounted on the staging at the breeching (or by
the furnace). A shed approximately 3x3x4 meters was constructed on the building floor
beneath the breeching, and the gas, pressure, and  temperature  lines were brought to the
equipment in the shed. There was  approximately 10 m of sample line between the probes
and the shed.
                                          55

-------
TEMP.
     H20
TEMP.
     H20
TEMP.
     H20
                 ANALYTICAL INSTRUMENTATION GASES, PRESSURE (API
                  Figure 14   Overall emissions measurements approach.
   STAGING
                      ANALYTICAL SHED
                                                                             6 Ipm
                                 BALLAST
                                  (45min)
                              (LATER REMOVED)
                                   BALLAST
                                   (15 min)
                               (LATER REMOVED)
                                 Figure 15   Sampling system.
                                            56

-------
     The original intent was to make possible the collection of the  H2O for analysis by
condensation and volume measurement at the end of a run. Similarly, particulate would be
collected in the cyclone, lines, knockout flasks, and first filter. Our initial runs showed that
the moisture content was quite low, and that only about  one-half of the H2O could be
collected by condensation. Consequently,  a separate sample  line — not cooled — was
strapped to  the  water-cooled gas sample probe and H2O determined by weight pickup on
1 cm x 10 cm silica gel traps connected to a conventional gas-sampling system.

     The three probes sampled gas constantly at a high enough flow rate so that we  could
always be assured of a representative incinerator sample, even though there was a relatively
large volume in the system. The combination of a lower than desired flow rate required to
prevent too high a pressure buildup on the  10-cm filter and the probe wall-effect problems
eliminated the possibility of obtaining representative particulate samples for %C analyses.

     Smaller pumps constantly pulled an aliquot  of the gas from each of the three filtered
combustion  gas streams for the instrumental measurements. In the  baseline series of tests,
gas chromatographs (GC) were used for the H2, CO2-O2-N2  analyses. One of the GC units
required 15 minutes for a full cycle; therefore the portion of the gas stream (0.1 Cpm) used
for these analyses was passed through a 4-£ container having a 45-minute time constant. In
this manner the analyzers could switch  between the gas stream every 15 minutes and obtain
a sample for analysis representative of the entire  time period preceding the last analysis of
that line. The more rapid CO and HC analyzers were cycled every 15 minutes, 5 minutes per
sample line.

     The results from the baseline studies showed that incinerator  conditions really  could
not be held  constant for  5 to 6 hours,  and  that observations would have to be made over a
shorter time (1 hour). Consequently, the ballast cans  in the gas sample lines were removed
and continuous CO2 and O2 analyzers  were substituted for the GC analyzer. Each sample
line was then analyzed for 5 minutes, giving a 15-minute cycle time for all three probes for
each analysis.

     The gas analysis system is shown schematically in Figure 16 as used in  the preferred
configuration with  continuous analyzers and without the ballast cans. The much greater
time variation in composition than expected places a clear preference on continuous and
rapid analysis,  and  we  strongly  recommend this approach over  the slower batch-type
analyses in all future studies.

     Basically the  instruments are simply  switched between the three probe gas lines by
means of the solenoids SI-S3 and S'l-S'3 and cycled between them every 15 minutes with 5
minutes on each probe line. The two-way solenoids allow continuous flow through the lines,
so that a representative sample of the incinerator at that moment is always present.

     The calibration gases were manually switched into the instruments for zero and span
checks at the beginning and end of each period of data collection.
                                          57

-------
                                                                    PROBE LINES
oo
            CALIBRATION
                GASES
              HiCO2,O2

              L0C02,02
                ZERO AIR
                PROPANE
1
>v
> 1
v

S1
\ v
1
V
j

•a
) S2
^ V
•) (

S'1
k /
S (
S'2
X.-..I
                                                                                    ANALYTICAL PUMPS
SOLENOID
 TIMING
 CIRCUIT
                                                                                  ^TS
                                                                                  S'3   ^.1    X
                                                          Figure 16  Gas analysis system

-------
     The starch CO/CO2 monitor ran continuously during periods of test activity monitor-
ing the concentration at point H in Figure 13.

     The pressure measurements were made with the S-type pitots on the gas-sampling
probes and various  other  static  pressure taps in the incinerator system, as indicated in
Figure 13.  The pressure-measuring lines were connected through a separate timer-controlled,
solenoid  switching manifold system,  as shown schematically in Figure 17.  A set of 10
measurements was made and recorded automatically every 10  minutes. The sequence of
measurements is given in Table 3. The recorded signal from the capacitance manometer was
calibrated for ± 1 cm H2O with an inclined manometer.
       Table 3.     VELOCITY AND STATIC PRESSURE MEASUREMENT LOCATIONS

Sequence            Function                         Probe
                                                    Probe Location'
    1
    2
    3
    4
    5
    6
    7
    8
    9
   10
Manometer Zero Check                    —
Velocity                            S-Pitot
Static Pressure*                       Static Tap
Velocity                            S-Pitot
Velocity                            S-Pitot
Static Pressure*                       Static Tap
Static Pressure*                       Static Tap
Static Pressure*                       Static Tap
Pressure Drop across Scrubber           Static Taps
Furnace Room — Outside Building        Static Taps
 Pressure Comparison
D
D
E
F
F
B or A**
C
GandH
Furnace Room
and Outside Building
 'Reference:  furnace room pressure
**Location changed from B to A for measurements on and after April 17, 1973.
t See Figure 13.
ANALYTICAL EQUIPMENT

Probe

     Considerable initial evaluation and trade-off consideration went into the final design of
the gas-sampling probe. The  first  experiments showed that  an uncooled 1.2-cm SS tube
could not physically withstand the 900-1000°C breeching temperatures and, more impor-
tantly, that  further reaction of the gases occurred in the hot sampling tube. When evaluated
                                          59

-------
                                                        TO PITOTS AND STATIC PRESSURE TAPE
o\
o
                          TO SOLENOID VALVES
                               FUNCTION
                               INDICATOR
                                 LIGHT
                                 PANEL
                                 TIMER
                               (CAM TYPE)
                                                                  4        cj>   4
4
                                   SOLENOID
                                    VALVES
                                                                                                 PURGE LINES
                                                                                     _Q—   (TO COMPRESSED AIR TANK)
                                                                  Q        Q  MANUALSHUTOFFVALVES
                                                                                ZEROING SOLENOID VALVE
CAPACITANCE
MANOMETER
                                                                      STRIP
                                                                      CHART
                                                                    RECORDER
                                              Figure 17   Static pressure and velocity (AP) system

-------
materials which might physically withstand the temperatures either had an excessively high
cost or a high probability of breakage. The EPA loaned us a water-cooled probe* for design
evaluation, and we found  that, using  this probe, we could quench the combustion gas
immediately after  it entered the probe, and that relatively inexpensive materials could
survive the incinerator environment without too much damage.

    Three probes were designed and  constructed as shown schematically  in Figure 18
(Section AA). This type probe essentially consists of a pair of concentric tubes with an inner
tube for  inlet cooling water. An additional internal tube was included for transmission of
the thermocouple leads and fastening the thermocouple.  A glass liner was used to decrease
catalyzed wall reactions of  the combustion gases. The first probes had a 90-deg sample inlet
so that samples, including  particulate,  could be obtained  parallel to the flow of the gas
stream. Two  problems were encountered with this design. The welds in the region of the
90-deg angle could  not withstand prolonged 1000°C temperatures. In fact two probes failed.
Furthermore, we  found that the impacting fly ash simply  caked over the gas  inlet and
caused sampling problems. Consequently, the ends of the probe were cut off at Section AA
and reconstructed as shown in  Section BB. Sampling  perpendicular to the gas stream
eliminated the fly ash caking problem and none of the modified probes failed.

    The S-type pitot tubes clamped to the probe worked well when clean, but, like the gas
inlet,  the upstream side  of the pitot  constantly  became  caked over and fused.  Reliable
pressure (AP) measurements could only be obtained with frequent (every hour) removal of
the probe and cleaning of the pitots.

    It was very important to maintain a high  flow of cooling water through the probes,
since even a very brief overheating caused probe sagging, linear breakage, and a run abort. In
a project such as this, even the simplest things become  problems. For example, we found
that high-quality garden hose used to supply the probes could not  be trusted and had to be
carefully  pressure-checked for burst strength.

    Temperature  measurements were  made with an Omega Pt/Pt 10% Rh thermocouple
housed in  a ceramic casing. The thermocouple  was held on  the  end of the  probe at
position 1 in Figure 18 as  shown in Figure 19.  The details of the radiation shield used to
protect the thermocouple from the incinerator  radiation and assure  accurate gas tempera-
tures are  also shown.

     Some measurements of "temperature only"  were made over the burning refuse bed,
while gas compositions were determined in the breeching. For these measurements, three
simpler water-cooled probes were made similar to  the gas probes, but they contained only
the line for the thermocouple and the radiation shield.
"Courtesy of Mr. James Dorsey, Control Systems Laboratory.
                                          61

-------
ON
to
                1.
                2.
                3.

                4.
                5.
                6.
                7.
                8.
                9.
               10.
               11.
               12.
                           	.LI	
                                                                                                                   co co
                                                                                                                   m m
                                                                                                                   O O
                                                                                                                      m
THERMOCOUPLE WELL - 3/8-INCH BWG20, 316 S.S.
COOLING WATER - 1/4-INCH TYPE L, COPPER TUBING
SAMPLE OUTLET - PYREX GLASS LINER, 5/8-INCH O.D.
               16 M.M., WITH 28/12 BALL JOINT
SAMPLE TUBE RECEIVER - 1/2-INCH SCHEDULE 10, 316 ELC-S.S.
SAMPLE INLET - 3/8-INCH SCHEDULE 80, 316 ELC-S.S.
S-TYPE PITOT TUBE - 3/8-INCH BWG 20, 316 S.S.
COOLING WATER INLET - 1/4-INCH SCHEDULE 40, 316 ELC-S.S.
COOLING WATER OUTLET - 1/4-INCH SCHEDULE 40, 316 ELC-S.S.
PROBE JACKET END PLATE - 316 ELC-S.S., 11 GAGE 2B FINISH
PROBE JACKET -  1 1/4-INCH SCHEDULE 10, 316 ELC-S.S.
MOISTURE SAMPLING TUBE- 1/4-INCH BWG20, 310S.S.
WORM-DRIVEN HOSE CLAMPS
co r= H oj
E! r- o m
Z S 2 "" ••
£°>53
10 —I > CO (/)
l>5°o
I O O 5 o
sgs§3

lll|i
3 o H 6 3
•  I CO Z O
  sl°;
  r= I o 3J
  H > oo >
  = H m Z
  S co <. O
  5>ls
  2l3l
    m .'
      m
                                                 Figure 18  Water-cooled sampling probe.

-------
           21/20.D.x0.049,304S.S,
            1 1/2 O.D.x 0.049, 304 S.S,
            _0.035 D.
             304 S.S. WIRE
       3/8 O.D. x 0.035 WALL, 304 S.S.


       FULL WELD



       316 S.S.


       .SET SCREW -2 PLACES, 90° APART


       3/4" D

       .	END OF COOLED PROBE
                                                        1/2 O.D.x 0.035, 304 S.S.
                 MODIFICATIONS

2 1/2 O.D. WALL INCREASED TO 0.065
1 1/2 O.D. WALL INCREASED TO 0.065
3/8 O.D. TUBE REPLACED WITH 1/4 INCH SCH 40, 316 S.S.
1/2 O.D. TUBE OMITTED
0.035 INCH WIRE REPLACED WITH 0.093 316 ELC S.S. ROD
WEDGE WAS ADDED FOR STRUCTURAL SUPPORT, 2 x 1 90° TRIANGLE
  MATERIAL-316 ELC S.S., 11 GAGE
        Figure 19   Thermocouple radiation shield
                       63

-------
Instrumentation

     The following  instrumentation  was used for  monitoring the concentration of the
species indicated:

     H2  —    Carle Model 8004 gas  chromatograph with automatic timer and sample
              injection.

     CO2,O2 - Carle Model 8004 gas chromatograph with automatic timer, sample, and
     N2, CO   column switch.

     CO2 -    Beckman Model 31 SB NDIR.

         -    MSA Model 300 LIRA NDIR.

     CO  -    MSA Model 300 LIRA NDIR (two).

     HC  -    Beckman Model 108A  FID.

     O2  -    Beckman Model 742 polarograph.

     Pressure — MKS Baratron capacitance gauge.

     Temperature —   L and N multipoint recorder.

     The main gas-probe sampling pumps were  made by  Cast  (Models 0322),  and the
analytical measurement pumps were made  by Metal Bellows Corp. (MB-21). Rockwell dry
gas meters were also used, and RAC inline aluminum filters were used for both locations. All
fittings were Swagelok, and  Nupro needle and toggle valves  were used. All components of
the gas analysis system were brass and  copper.

Photographs of Equipment

     Figures 20 through  23 are photographs of the various pieces of equipment in the
analytical shed. Figure 20 is the three main gas sampling systems as represented earlier in
Figure 15. Figure 21 shows the multipoint temperature recorder and the original three-pitot
pressure-measurement system. The pressure system was  later expanded to the  10-point
measurement manifold. Figure 22 is a view including the NDIR CO and FID HC analyzers
and several of the recorders. Figure 23 shows the two original gas chromatograph analyzers
(one was later replaced  with  continuous CO2  and  O2  analyzers) and  the CO NDIR
analyzers.
                                       64

-------
Figure 20   Gas sampling lines, pumps and meters.

-------
Figure 21   Recording temperature and pressure systems.
                          66

-------
Figure 22   NDIR and FID analyzers and recorders.
                      67

-------
Figure 23   Gas chromatograph and NDIR analyzers.
                        68

-------
DATA REDUCTION

     The data reduction was primarily a three-step operation. The form of the data was first
changed from percent displacement on a strip chart recorder to  a tabular listing of data in
appropriate units. In the second step, the data for a particular run  were averaged so that each
run could be characterized by a simple set of output parameters.  Finally, the validity of the
output parameters was checked by mass and energy balances.

Data Recording and Averaging

     The first and second steps were purely mechanical manipulations of the recorded data.
In reading data points from a continuous strip chart recorder,  five-minute averages were
taken at five-minute intervals. This particular method was chosen  because of the five-minute
sample time for each of the breech zones. The raw averaged data were then applied to the
appropriate calibration curve to establish the data in desired units of temperature, concen-
tration, or velocity. The moisture content data were recorded in 15-minute intervals as
15-minute averages; this timing corresponded to one complete  cycle of the breech zone
sampling system. To calculate the set of output parameters for each run, the tabulated data
were time-averaged for the entire run.  The output parameters are listed below:

     •   Temperature
         Breech zones 1, 2, and 3 and the average temperature over the entire breech

     •   Flow rate
         Velocity in breech zones 1,2, and 3
         Total molar flow rate

     •   Composition of exhaust gas
         Breech zones 1, 2, and 3 - CO2, CO, CH4, H2O, and 02, and the average
         molecular weight
         For total breech - CO, CH4, O2, N2, CO2, and H2 O

     •   Heat output of system

These output parameters were used for the remainder of the analysis.

Data Checks — Mass and Energy Balances

     The  difficulties  encountered in obtaining  valid data from an  operating municipal
incinerator are  well known. The conditions under which sampling must take place are hot
(1200 to 2000°F),  dirty (molten fly ash, heavy organics, high moisture content), and not
chemically constant from second to  second or, on the  average,  from hour to hour. These
difficulties do  not  preclude the possibility of obtaining valid data,  but they  do  place
increased significance on material and energy balances as a means of verifying the internal
                                         69

-------
consistency of the data. Although we accept that mass and energy balances are  not the
necessary and sufficient criteria for proving the validity of data, reasonably good balances
do tend to increase the credibility of the data and, hence, our confidence in it.

     As a rule, then, whenever balances can be closed, the data are considered valid. For
experiments in which the balances did not close, the data were rejected. In the latter cases,
we attempted to explain the most probable source of error or inconsistency.

The Mass Balances

     The overall mass balance over the incinerator was impossible to obtain because a large
fraction of the air entering the furnace was not metered  (leakage). The amount of leakage
was  determined  by the  difference between the  exiting gases and  the forced  air and
combustion-generated gases.

     Carbon — The only important source of carbon in the system is the  refuse feed. It is
assumed that the average refuse at Newton is 50% combustible, and that these combustibles
have an empirical formula of CHj 585O0  675  based upon previous data for Newton  [8].
Carbon leaves the system as CO2, CO, various  hydrocarbons, soot, and unburned material
which  falls out of the incinerator as ash. For a typical run the distribution  of carbon among
these various forms is shown in Table 4.

                        Table 4. CARBON  DISTRIBUTION PROFILE
      Component
 Pounds
(mols C/hr)
Percent
     Carbon In
      Refuse
      CO2 in air
  415.71
  413.35
    2.36
100
 99.43
  0.57
     Carbon Out
      C02
      CO
      Hydrocarbons
      Soot
      Ash
  415.71
  410.31
    1.49
    0.56
    0.21*
    4.14*
100
 98.46
  0.36
  0.13
  0.05
  1.00
     'Arthur D. Little, Inc., estimates.
                                          70

-------
     Because carbon dioxide represents such a large percentage of the total "carbon out," in
the remainder of the  analysis  we have ignored the contributions from CO, hydrocarbons,
soot, and the ash. By  the same reasoning only the "carbon in" as refuse will be utilized in
further carbon balances.

     For these specific data, the combustion of refuse is represented by the following formula:

CH1.585 °o.625 • 0.655 H2O+ 1.084 02 +4.078 N2—>CO^ + 1.448 H2O+ 4.078 N2

                                                                             (44)
This relationship suggests that  1  pound  mol of CO2 should be produced for every 23.585
pounds  of combustible (or 47.17 pounds of refuse).  Because of the uncertainty  and
variations in the bulk density of refuse, as well as the voids which are present in the feed
hopper, it is difficult to calculate  the exact feed rate. However, for the purposes of the
material balances, feed rates were calculated assuming a density of 13 lb/ft3 and that there
were no voids in the feed  hopper.  It can be seen from Table 4 that the carbon in and the
carbon  out  are in  reasonable agreement.  From  the  information in this table  one  can
calculate that the average  amount  of combustible fed into the furnace was 51.5% of the
refuse (p = 13 lb/ft3). For  the  purposes of the current  work, this agreement is adequate to
verify that the techniques and apparatus are appropriate.

     Hydrogen  — Hydrogen enters  the system as water vapor in the air, as moisture in the
refuse, as bound hydrogen  in the refuse, and as water evaporated  from the ash quench
system. The hydrogen leaves the system as water vapor and hydrocarbons in the exhaust
gases and as moisture and  bound hydrogen in the unburned combustibles. A  hydrogen
balance for a typical run is shown in Table 5. As in the case of the carbon balance, the small
contributors have been ignored; that is, hydrocarbons and hydrogen in ash either bound or
as moisture and the water vapor in the air from the ash quench system. The "hydrogen out"
is all measured as water vapor. The "hydrogen in" is calculated from the "carbon dioxide
out" (which has previously been related to amount of refuse fed). The empirical relationship
states that the mols of water released  is equal to  1.448 times the mols of CO2 released. As
can  be seen, the water balance (which  implies the hydrogen balance) is good  enough to
justify the use of the  data. In the current study we found the average moisture content of
refuse to be 24.6%, as compared to the assumed value of 25%.
                                         71

-------
                    Table 5. HYDROGEN BALANCE FOR TYPICAL RUN

                                             Pound
      Component                            (mols H/hr)                        Percent

     Hydrogen In                              1373.24                        100
      HasH2OinAir                           122.90                          8.94
      H as H20 in Refuse                         541.58                         39.45
      H (bound) in Refuse                        655.16                         47.71
      H as H20 from Ash Quench*                   53.60                          3.9

     Hydrogen Out                             1373.24                        100
      H20 in Gases                             1359.13                         98.97
      H as Hydrocarbons                           2.13                          0.16
      H20 in Ash                                 5.42                          0.39
      H (bound) in Ash                            6.56                          0.48
'Arthur D. Little, Inc., estimate.
     Oxygen —  The oxygen in the system comes from the atmospheric oxygen and water
vapor, the moisture in the refuse, the water evaporated from the ash quench, and the bound
oxygen in  the  refuse  combustibles.  The oxygen leaves the incinerator as atmospheric
oxygen, carbon  dioxide, carbon monoxide, and water vapor. A typical  run has an oxygen
balance similar to that presented in Table 6.
                       Table 6. OXYGEN BALANCE IN TYPICAL RUN

                                             Pound
     Component                             (molsO/hr)                        Percent

     Oxygen In                                3673.65                        100
      O as O2 in Air                            3056.27                         83.20
      OasH2OinAir                             61.45                          1.67
      0 as H20 in Refuse                         270.79                          7.37
      0 as H2O from Ash Quench*                  26.80                          0.73
      0 as Bound Oxygen in Refuse                 258.34                          7.03

     Oxygen Out                               3673.65                        100
      O as O2 in Gases                          2166.68                         58.98
      OasH2Oin Gases                          679.56                         18.50
      0 as C02 in Gases                          820.62                         22.34
      0 as CO in Gases                             1.49                          0.04
      OasH2OinAsh*                            2.71                          0.07
      O as Bound Oxygen in Ash*                    2.59                          0.07
'Arthur D. Little, Inc., estimate.
                                          72

-------
The small contributors - CO, oxygen in ash as moisture and bound water, and water from
ash quench — are not considered when solving the oxygen material balance.

     A summary of the average oxygen balances is given in Table 6. As can be seen, these
balances are adequate for the current work.

The Energy Balance

     The energy balance  over the incinerator has been used in the same way as the mass
balances  to verify the experimental techniques and the analytical apparatus.  Essentially,
there are two ways of conducting the  energy  balance. The first  is to assume a constant
heating value  for refuse  and check the heat actually liberated with  the theoretical heat
available. The other is to compare the heat liberated per unit of combustion product
formed. The latter is the  chosen course because it does not presume knowledge of the heat
content of the raw refuse feed or the feed rate.

     It is known that about 5650 Btu's are released per pound mol of CO2 liberated in the
combustion of many  organic materials. Therefore, one could conduct the heat balance by
knowing  the mass flow, composition, temperature, and heat capacity of the gas stream. On the
other hand, the same  heat balance could be accomplished knowing only the fraction of the
flue gas which is CO2 and the temperature. If we assume that the heat capacity of the flue
gas stream does not change with composition, then the temperature of the flue gases will be
a linear function of the fraction of CO2 (Appendix D, Eq.D-31). Furthermore, the theoretical
line passes through the adiabatic flame temperature at the stoichiometric CO2 concentration
and through ambient temperature at a CO2 concentration  equal to zero. Figure  24 shows
the data  on such a plot  of temperature vs CO2  concentration. Because the system is not
adiabatic, it is reasonable to  assume that the dotted curve represents the kind of deviation
from the theoretical that we might expect, that is, the largest deviation at higher tempera-
tures and no  deviation at ambient. Rotation  of the line  about  the  average temperature
would also be expected as a result of radiation. From the slope of a line through the plotted
points, which is about .004% C02/°F (or 260°F/%CO2), we  have calculated that the heating
value of the refuse is about 4650 Btu/lb based upon Eq. (D-31) of Appendix D.
                                         73

-------
1   -
             750
1000
  1250
TEMP.,°F
1500
   O ZONE 1
   A ZONE 2
   D ZONE 3
1750
2000
                         Figure 24  CO2 vs. temperature.
                                 74

-------
                                    SECTION VII

                                  TEST RESULTS

     The  strategy we espoused for controlling  combustible emissions  in Section V  was
predicated on two separate mixing concepts:

     •    Macroscale mixing to achieve uniformity throughout the furnace; and

     •    Microscale mixing to  promote burnout of combustibles at the molecular
          level.

     The  equations that we presented as the basis for each of the two mixing models are
verified in this section. The data were taken at an existing municipal incinerator which was
retrofitted with overfire air jets for this program.

     The  test equipment and procedures were discussed in Section VI. In this Section, we
will  describe the  specific tests that we carried out and discuss the results of those tests. A
compilation of the test data is given in Appendix A.

MACROSCALE MIXING MODEL

     Overfire jets can be used  to  achieve uniform  mixing throughout  the furnace.  The
macroscale model proposes a design basis for overfire air jets based upon:

     •    Ivanov's equation for jet penetration, and

     •    Use  of Bernoulli's equation to characterize the  flow patterns within the
          furnace.

     The  method for calculating the  penetration of an overtired jet into the furnace was
suggested  by Ivanov and has been reviewed in Section V. The design equation (Eq. (3)) is
repeated below:
                                           i—
                                                   d:Uj
                                                 —L-L                          (3)
     The equation  is dependent upon  the  jet diameter, its velocity, and  the crossflow
velocity of gas in the furnace. In a series of jet tests the penetration distance was estimated
visually and was compared with the calculated value for penetration. Figure 25 shows a plot
of penetration vs jet velocity for 5-inch diameter jets firing into a crossflow velocity field of
15  ft/sec. The crossflow velocity was calculated  using the fluid-flow equations discussed
previously. Clearly, the Ivanov relation  is adequate for estimating jet penetration, and the
deviation from the theoretical relationship is well within experimental error.
                                         75

-------
o
{= t; 4
tJ"
          dJ
             = 1.6d,u,
=  5"
=  15fps
            p\
                 =  1-8
              10
       20
30
40
50
60
70
80
90
100
                          AVERAGE JET VELOCITY, Uj.ft/sec

                           Figure 25  Penetration vs. jet velocity.

 FLU ID-FLOW PATTERNS

     The fluid-flow model developed in the mixing study is based on the premise that
 Bernoulli's equation can be applied along the streamline of flow in an incinerator so that
 both the residence time and the crossflow velocity at the level of the jets can be calculated.

     During  the  tests,  velocity  and temperature were measured  in  the  breech of the
 incinerator. The  profiles  clearly showed the stability of the hot zone at the top of the
 breech, but the data did  not  show large difference between the "hot" and "cold"  zone
 velocities. The equations derived for a sealed furnace [Eq. (8)] could not correlate the data,
 as witnessed  by the random appearance of a plot of u2  vs 2g(z-z0) [(TH/TC) -1 ]. On the
 other  hand,  Eq.  (9) derived  for unsealed furnaces  could  correlate the data reasonably
 well. A  plot  of  u2  vs  [(T/T0)-1]  where T0  is the ambient temperature and  Tis the
                                         76

-------
temperature of either the hot zone or cold zone is shown in Figure 26. The plot should be
linear if Eq. (9) repeated here is applicable to incinerators:
                 Hot Zone    ,        v                    Cold Zone
     With the  exception of a few  spurious points, the agreement  with the theoretical
relationship  is excellent.  The solid line in the figure represents the best fit through the data
points corresponding to a furnace height of 12 feet and an initial velocity of 18.4 ft/sec. The
dashed line represents the curve passing through the origin, so that the furnace height is 13.7
feet, while the initial velocity is zero. Because of the gross nature of the flow patterns within
the furnace, the analytical distinction between the  height of 12 feet and 13.7 feet is not
possible. Either curve could be justified as confirmation of the Bernoulli approach; neither is
so well defined that it may be used to predict initial velocities.

     Two conclusions can be drawn from this analysis:

     1)   Bernoulli's equation can, indeed, be used for design purposes in determining
          exit velocities and approximate furnace residence times; and

     2)   In a furnace which has considerable air leakage, both the hot and cold zones
          will accelerate  according to the Bernoulli relation. The driving force for the
          acceleration  is the buoyancy  of gases within the  furnace  as  compared to
          outside the furnace.

     Verification of Eq. (10) of Section V, which describes the difference in velocity of the
hot zone vs the cold zone, was not possible. The accuracy of the data does not warrant such
fine tuning of the fluid-flow model.

MICROSCALE MIXING  MODEL

     The theories presented to characterize the microscale mixing introduced the use of
Hottel's unmixedness  factor and  Corrsin's equation as fundamental relationships  charac-
terizing  the three T's on a  microscopic scale. Hottel's and  Corrsin's concepts  have  been
verified in controlled situations, but because of the large variations in incineration data it is
unrealistic to hope for conclusive evidence that will provide exact quantitative relations. The
data are sufficiently accurate, however, to demonstrate  the consistency of the data with the
two concepts.

The Unmixedness Factor

     In  Section V, the relationship between the unmixedness factor suggested by Hottel (4)
and the ratio  $ of available  oxygen to  oxygen required to complete combustion  was
                                          77

-------
oo
                     3.0
                                                 where
                                                      T.=70 F, us 18.4 ft /sec
                                                       a          LI



                                                      z-z0  =12ft.
                                                                                                         Tas70°F,  u0 =
                                                                                                          -z)^ 13.7ft.
                     1.0
                        1500
1750               2000               2250





                                u,2  [ft2/sec2]






            Figure 26    Confirmation of Bernoulli's equation.
                                                                                                    2500
2750
3000

-------
presented. The relation is derived in Appendix F. The procedure for calculating the value of
the unmixedness factor is outlined in Appendix B. Supporting data required to characterize
the refuse bed offgas is presented in Appendix E.

     The unmixedness factor can be calculated for each gas sample that is taken. Its value is
dependent only upon the gas analyses and the empirically determined value for CO/CO2.  The
experimental value  for yV^/Ay is determined  from Figure 3, based on  the value of $
[(available oxygen)/(oxygen required  to complete_combustion)], determined from the gas
analysis. The value of the unmixedness factor, y c'2, is calculated from the following equa-
tion which is based upon Eq. (B-12):
                                 77 \     /   o/*r*r\   \
                                                                               (45)
     In Appendix B, the quantity ys is shown to be equal to 0.78 when CO/CO2 is equal to
0.5. The composition of the combustibles and combustion products (without entrained air)
is also given.

     In Table 7, the experimental value for y c'2 is given for each breech zone and each of
the 28 jet tests.

     To compare the experimental value for unmixedness factor with the theoretical one,
the following data are required:

     •   original CO/CO2 off the bed, and
     •   relationship between temperature and CO2 in the breech.

     In these calculations, a value of 0.5 for CO/CO2 has been assumed.

     The plot of temperature vs %CO2 |w was given previously in Figure 24.

     The data can be represented by the linear relationship:
                            %CO I
                                2 w   =(Temp-500°F)/1500°F                 (46)
                              10%

     The calculation procedure is as follows:

     1.  For  a given temperature, calculate %CO2 |w  from the equation given above;

     2.
                                            ^

     3.    For a given value of yc'2 , calculate lye'2 /^yy ;
                                         79

-------
                    Table 7.  SUMMARY OF UNMIXEDNESS FACTORS
Zone I
                                             Zonal!
                                                                                     Zone III
Run#
22-1
222
22-3
23-1
23-2
23-3
23-4
24-1
24-2
24-3
24-4
OO 25-1
° 25-2
25-3
25-4
31-1
31-2
31-3
32-1
32-2
32-3
33-1
33-2
33-3
34-1
34-2
34-3
34-4
Emission
Factor
1
1
1
127
87
330
252
144
60
55
409
652
976
106
91
157
46
26
243
207
235
275
86
505
272
67
61
54
Temp.
1721
1508
1700
1715
1437
1717
1421
1647
1694
1551
1730
1900
1796
1400
1523
1750
1507
1525
1825
1683
1698
1777
1666
1753
1793
1554
1497
1597
V^L
<0.12
<0.14
<0.12
0.145
0.165
0.160
0.190
0.150
0.135
0.150
0.165
0.155
0.180
0.170
0.160
0.145
0.170
0.145
0.140
0.155
0.155
0.150
0.140
0.150
0.150
0.150
0.155
0.140
V^w
.062
.076
.065
0.116
0.157
0.149
0.200
0.174
0.115
0.126
0.132
0.099
0.193
0.147
0.135
0.093
0.135
0.125
0.139
0.130
0.161
0.113
0.130
0.159
0.138
0.129
0.138
0.118
^
_
_
-
0.029
0.008
0.011
- 0.010
- 0.024
0.020
0.024
0.033
0.056
- 0.013
0.023
0.025
0.052
0.035
0.020
0.001
0.025
- 0.006
0.037
0.010
- 0.009
0.012
0.021
0.017
0.022
Emission
Factor
1015
583
244
677
909
3705
2707
535
18
24
24
499
462
568
1387
267
442
572
49
182
418
550
416
545
320
502
596
412
Temp.
1446
1387
1458
1446
1125
1350
1158
1385
1470
1341
1426
1696
1587
1246
1131
1543
1261
1308
1707
1667
1609
1545
1485
1585
1478
1417
1367
1481
V^
0.222
0.210
0.180
0.210
0.250
0.290
0.290
0.210
0.150
0.160
0.150
0.170
0.185
0.225
0.265
0.175
0.215
0.220
0.130
0.155
0.180
0.190
0.190
0.190
0.185
0.205
0.215
0.190
^u-
0.237
0.223
0.172
0.223
0.244
0.312
0.287
0.210
0.149
0.159
0.160
0.167
0.191
0.219
0.265
6.173
0.221
0.214
0.139
0.157
0.198
0.202
0.193
0.195
0.185
0.214
0.230
0.196
&
- 0.015
- 0.013
0.008
- 0.013
0.006
- 0.022
0.003
0
0.001
0.001
- 0.010
0.003
- 0.006
0.006
0
0.002
- 0.006
0.006
- 0.009
- 0.002
- 0.018
- 0.012
- 0.003
- 0.005
0
- 0.009
- 0.015
- 0.006
Emission
Factor
1314
274
931
460
979
12.162
7953
1102
1633
2461
2508
823
1242
2060
5141
823
193
142
1403
84
1042
2449
2012
3340
2484
2594
2862
3689
Temp.
1029
1004
1100
1046
737
871
856
892
1152
949
1071
1283
1225
875
742
1193
836
1025
1381
1408
1344
1251
1240
1306
1150
1106
1023
1115
1C '«xpt.
0.270
0.225
0.250
0.235
0.280
70.40
0.370
0.270
0.270
0.300
0.295
0.230
0.250
0.295
0.350
0.240
0.230
0.210
0.240
0.170
0.230
0.280
0.270
0.290
0.290
0.295
0.305
0.320
v^
- 0.280
0.240
0.278
0.239
0.241
0.350
0.338
0.262
0.296
0.306
0.308
0.304
0.270
0.297
0.325
0.265
0.241
0.231
0.290
0.211
0.272
0.304
0.293
0.311
0322
0.312
0.315
0.325
^
- 0.010
- 0.015
- 0.028
- 0.004
0.039
0.050
0.032
0.008
- 0.026
- 0.006
- 0.013
- 0.074
- 0.020
- 0.002
0.025
- 0.025
- 0.01 1
- 0.021
-0.050
- 0.041
- 0.042
- 0.024
- 0.023
- 0.021
- 0.032
- 0.017
- 0.010
- 0.005

-------
     4.   Determine $ from Figure 3 ;

     5.   Calculate %O2  from the material balance  equation (Eq. D-26 of Appen-
         dix D):
                                 %o2  t  %co2i      _
                                 21%     15.1%

     6.   Calculate O2 required, according to the relation:
     7.   Calculate the emission factor according to the equation:

                       CO(PPm)|12% = (02req)


Note that for refuse:
                                %co2lw  =%co2id                            (47)
                                  15.1%      19.5%

Using the above steps, we calculated  the data shown in Figures 9,  10, and 11. The
procedure shows that emission factors  can be calculated  if the temperature and
unmixedness factor are specified and if the original composition of the offgas is
known.

     The theoretical value for y c'2 has been determined from Figure 8 for each of the
jet tests, and these values are tabulated in Table 7.

     A careful examination of  the data reveals that  in all cases, even in the quench
zone, high emissions are reflected in high unmixedness factors. The agreement between
experimental and theoretical values  overall is 0.003 or within  1%  of the expected
value. The agreement is best in  Zone II where the data from  Figure 24 are best fitted
by Eq. (46).

     The parameter (CO/CO2)0 has no effect on the relative values of theoretical and
experimental unmixedness factors, since the ratio was  used to  calculate both. However.
a change in  the value of (CO/CO2) 0 results in a change in the absolute value of^/c^
Hence, to attempt to place any  weight on the actual numerical values in Table 7 would
be stretching the validity of the data.  With respect to the validity of the concept, the
correlation of the data is remarkably well fitted to the unmixedness factor concept.
Without this characterization factor, the more sophisticated advances in the theoretical
understanding of incinerator mixing do not readily follow-
                                         81

-------
Corrsin's Equation
     The verification of Corrsin's Equation  could be based upon the defining equation,
(Eq. (21)):
                                                                                    (21)
or alternatively upon the comparative equation (Eq. (40)):
                                  y
                             1 +
                                Ayb\%C02|s
                            d-x)
                                                                                    (40)
In Appendix  E,  tests are described wherein offgas compositions were  measured  directly
over the refuse bed. From these data the initial value of the unmixedness factor ^/ c7* was
calculated. The calculated values for the two complete bed tests are shown in Table 8. The
scatter in the data was so great that no meaningful conclusions were possible.
                     Table 8. SUMMARY OF TEST RESULTS OVER
                                       Sample Sort
Bed Test 33

C02
02
CO
H2
HC
H20(>)
H20/C02
CO/C02
Kwg
Temperature

Bed Test 2

C02
02
CO
H2
HC
H20
H20/C02
C0/C02
^wg
Temperature
 5.93
13.4
 3.2
 1.72
 1.16
11.51
 1.94
 0.54
 3.61
1835
13.3
 6.0
 3.35
 2.7
 1.94
12.9
 0.97
 0.25
 1.20
1787
 1.31
18.9
 0.26
 0.21
 0.01
 2.15
 1.64
 0.20
 2.03
1905
 6.4
14.2
 1.00
 0.77
 0.34
19.4
 3.03
 0.16
 3.94
1758
 5.8
14.6
 0.13
 0.09
 0.03
 8.84
 1.52
 0.02
 2.2
1928
12.0
 8.6
 0.26
 0.19
 0.39
26.4
 1.70
 0.02
 2.33
1811
Average

 4.35
15.63
 1.20
 0.67
 0.40
 7.50
 1.72
 0.28
 3.1
1889
10.6
 9.6
 1.54
 1.22
 0.89
17.57
 1.66
 0.15
 2.09
1785
1.  Complete sets of data are limited due to experimental difficulties directly over the bed.
2.  Calculated from material balance.
                                           82

-------
     In pursuing the alternative approach to verification of the Corrsin relation, note that
neither the parameter  x nor the parameter (j3t)b are well known or clearly defined. The
value for x could be defined by Eq. (28) in Section V, based upon the ratio %CO2 |j/%C02 Ib •
One might also  define x based  on temperature by  substituting Eq. (46) into Eq. (28) of
Section V to yield:
                                            (T: - 500)
                                        --                                     (48>
Eq. (28)
0.284
0.260
0.163
0.139
Eq. (47)
0.194
0.180
0.087
0.019
Eq. (30)
0.290
0.294
0.277
0.256
     Finally, one might  also define x based upon flow rates (Eq. (30)). For example, the
different calculated  values for x in each of the test conditions of Test 31  are shown in
Table 9 to be in relatively good agreement. The values for Test 33 are grossly apart.

                     Table 9.  COMPARISON OF ESTIMATES FOR (1-x)

                                _ Basis for (1-x)
              Test No.
              31-2
              31-3
              33-1
              33-3

     The reasons for the discrepancies in the above values for x are:

     •   Errors in analytical data,
     •    Changes in total flow rates into the furnace, and
     •   Changes in composition or heating value of refuse.

To determine  the  effect  that  different values for x might have on the  resulting
calculations, consider the calculations summarized in Table 10.  In these calculations,
based upon Tests 31-1 (base condition) and 31-2 (jet condition),  x is chosen arbitrarily
and Ay is defined by the  equation:
                                            %C02|b

                                                                                 (49)
Note that as the value for (1-x) decreases, the value for the experimentally determined
unmixedness factor \/c^ also decreases (due to the change in the calculated value for
Ay, according to Eq. (49) above), and the predicted value for\/crT increases, primarily
due to the  decrease in the value of (|3t)j. The  result is the rapid convergence of the
experimental and predicted values for unmixedness factor to the extent that agreement
with the Corrsin relation can be shown.

     In the first set of calculations (3t  was assumed to be equal to 0.25. On the other
hand,  the  value of ()3t) is not  clearly defined,  because it depends on the arbitrary
                                         83

-------
                                      Table 10. CALCULATION RESULTS USING DIFFERENT VALUES FOR x
oo

Test
31-1
31-2


31-1
31-2


31-1
31-2



(1-x)
0
0.284
0.150
0.100
0
0.284
0.150
0.100
0
0.284
0.150
0.100

*
445
3339
3339
3339
445
3339
3339
3339
445
3339
3339
3339
From Data
f(*>
0.455
0.370
0.370
0.370
0.455
0.370
0.370
0.370
0.455
0.370
0.370
0.370
(Appendix
Ay(x)
0.207
0.371
0.294
0.265
0.207
0.371
0.294
0.265
0.207
0.361
0.294
0.265
B)
^
0.094
0.137
0.109
0.098
0.094
0.137
0.109
0.098
0.094
0.137
0.109
0.098
From Theory Application of Eq. (41 )
*0
86
203
138
119
150
354
241.
208
311
734
500
431
«*o»
0.584
0.500
0.535
0.550
0.528
0.465
0.490
0.500
0.477
0.425
0.445
0.455
V^0 1
0.121
0.186
0.157
0.146
0.109
0.173
0.144
0.133
0.099
0.154
0.131
0.121
+ ys (%C02/b) ,,._
Ay %co2/s
1.0
1.791
1.418
1.278
1.0
1.791
1.418
1.278
1.0
1.791
1.418
1.278
:) f(*j0)/f(4>b0>
1.0
0.856
0.916
0.942
1.0
0.881
0.928
0.947
1.0
0.891
0.933
0.954
*
0.250
0.745
0.420
0.321
0.150
0.447
0.252
0.193
0.050
0.149
0.084
0.064

N/c"
0.094
0.088
0.103
0.105
0.094
0.110
0.112
0.109
0.094
0.136
0.120
0.113
A%
—
-35.8
-5.5
+7.6
—
-19.7
+2.8
+11.2
—
-0.7
+10.1
+15.3

-------
assumption of "characteristics," lengths, and also upon assumed relationships between
residence time, furnace volume, and  flow rates (Eq.  (26)). In the subsequent sets of
calculations in Table 11, the effect of different values of£pt)b is demonstrated. As ()3t)
decreases,  the experimentally determined value for yc7 remains constant,  but the
predicted value, based  upon  Eq.  (40), increases.  At a value for (j3t)b of 0.05, the
experimental  and  predicted  values for  unmixedness  factor agree (based  upon  x
calculated from carbon dioxide ratios).

     The  other variables in Eq.  (40), such as the assumed  values for CO/CO2 and
(Af/Aj), are much less important and have not been considered in detail.
     Table 1 1 shows the relative agreement between yc   determined from the stack
gas analysis (based upon Ay calculated  from Eq. (49))  and the predicted value for
ye'2 , based upon Eq. (40). Note that for a given value of x, there is a value for (|3t)b
which will produce close agreement between the twoyc77 values. Likewise, for a given
value for ((3t)b ,  there exists a value for x which will also produce a close agreement
between the experimental and theoretically predicted values for yc72". In Figure 27, the
value of (1-x) is shown as a function of the value of (0t)b  required to reconcile the test
data for Tests 31-1  and 31-2 with the values calculated  from Eq. (40). The range of
mutually consistent values for (1-x) and  (j3t)b falls within the accuracy of the experi-
mental data. Because of the uncertainty associated with both x and j3t, one cannot
conclude that the validity of the Corrsin theory is "proved," but only that the data are
consistent  with  the  theory  to the  extent  that  the results of the tests could be
reasonably explained. Similar analyses could be made on each of the other series of
tests with some notable exceptions. In Test 32 the addition of jet air caused an increase
in CO2 in Zone 1 . A possible explanation for this is that the jet  air was introduced at a
much  higher  static pressure  than was  the air which leaks into the furnace.  The
introduction of jet air could change the static pressure pattern within the furnace to
the extent that  one mole of jet air displaces more  than one mole of leakage air.  The
result  was a decrease in the total amount of air  entering the furnace  and  a  corre-
sponding increase in the CO2  content. In this particular case then, the emissions are
not dependent so much upon the rate of decay as upon the basic change in combustion
conditions within the furnace itself.

     In Test 33-2 the furnace temperature was intentionally held constant. To achieve
this, it was necessary to  cut down on the underfire  air to the furnace. In this case also,
the poor test results resulted from significant changes in the combustion conditions
within the furnace rather than from the effects of turbulent mixing.

     Finally, reductions in combustible emissions in Zones 2 and 3, of the breech were
rarely seen. We propose two reasons for this phenomenon:

     •   The gases in Zones  2 and  3  emanate  from  portions of the  furnace not
          covered by overfire mixing jets, and
                                         85

-------
Table 11.  RELATIVE AGREEMENT BETWEEN
         ANDv^)calc.
                                                                  expt
                0.284
                0.150
                0.100
             0.25
             -35.8
             -5.5
              +7.6
 0.15
-19.7
 +2.8
+11.2
 0.05
 -0.7
+10.1
+15.3
    0.3
    0.2
x
    0.1
                    0.05
               0.10
                                                  0.15
                                            0.20
                             0.25
                                Figure 27  (1-x) vs. (0t)b.
                                         86

-------
     •   The temperatures in these zones are lower than in Zone 1 to the extent that
         a significant portion of the concentration distribution could extend into the
         quench zone. Unfortunately, the installation of overfire mixing jets in these
         colder portions  of the  furnace would not  be effective since the positive
         effect  of the increase in mixing intensity would be offset by the negative
         effects of extending  an even greater portion of the concentration distribu-
         tion into the quench zone.

RESULTS OF PARTICULATE TESTS

     As a complementary portion of the test program, pollutant emissions were measured in
the incinerator  stack for particulates, HC1, SOX, and NOX.  The particulates were analyzed
for trace metal content and polynuclear aromatic hydrocarbons.

     The emissions were measured using the standard  particulate sampling techniques (also
called Method 5) as  described in the Federal Register of December 23, 1971. Measurements
were made during the time  of the first five BASE tests. The gaseous species (HC1, SOX,
NOX) were measured in the duct leading to the base of the stack using standard impinger
collection techniques. Particulates were collected at a  point in the stack 60 ft. (18m) above
the foundation of the stack. Pertinent stack data at the sampling point follow:

              stack O.D.                  20 ft (6.1m)
              stack I.D.                   12 ft (3.6m)
              diameters upstream          10
              diameters downstream       3.4

     Because of the location of the  existing sampling ports, measurements were made by
traversing the stack on one diameter from both sides; a total of 40 traverse points were used.
The location of these points in each radius is given in Table 12.

     Results of the gaseous pollutant measurements are given in Table 13. The data are all
normalized to  12% CO2. The values are expected to vary with refuse composition as
observed, but there is a very wide range in the HC1 values from 220 ppm.

     The basic  stack particulate and sampling data are given in Table 14. The particulate
data are summarized in Table 15 on the basis of 12% CO2. The amount of particulate below
the 5- to 10-jz cyclone cutoff and collected on the filter is  83%, 80%, 84%, and 78% in test
samples 2-5, respectively.
                                        87

-------
     Table 12.  SAMPLING POINTS PER RADIUS
  Point Number          Distance from Inner Wall
        1                    5ft  0-5/8 in
        2                    4'   4-3/8"
        3                    3'  10-1/8"
        4                    3'   5-7/8"
        5                    3'   1-7/8"
        6                    2'  10-3/8"
        7                    2'   7"
        8                    2'   4"
        9                    2'   1-1/4"
       10                    1'  10"
       11                    V   8"
       12                    1'   5-3/8"
       13                    1'   4-1/8"
       14                    1'   0-7/8"
       15                       10-5/8"
       16                        8-5/8"
       17                        6-5/8"
       18                        4-1/2"
       19                        2-3/4"
       20                          7/8"
   Table 13.  GASEOUS POLLUTANT EMISSIONS
                   (at12%CO2)


              	Concentration (ppm) for Test

Species         1         234
NOX            130      85    220    125    170
SOX           1,170     530    240    970     38
HCI              28     220     81      8      2
                       88

-------
                    Table 14.  STACK PARTICULATE DATA

                                                      Test
Volume Samples (m3)
Stack Velocity (ft/sec)
Stack Temperature (°F)
Stack Static Pressure (in H20)
Moisture (%)
C02(%)

Particulate Weights (g)
   Probe + Cyclone
   Filter
   Aqueous Residue3
   Organic Extract8

a.  of impingers
2
4.63
25
389
0.02
21
2.7
0.127
0.634
0.161
0.112
3
2.94
.14
450
0.02
33
4.1
0.178
0.172
0.108
0.103
4
2.12
15
380
0.02
28
3.1
0.044
0.225
0.070
0.018
5
3.10
14
470
0.01
30
5.0
0.160
0.580
0.045
0.032
Portion
                    Table 15.   PARTICULATE ANALYSES
                                 (at12%C02)
                                          Particulate Weight (g) for Test
Probe + Cyclone
Filter'

Aqueous Residue
Organic Extract

Total Probe + Cyclone
   + Filter
0.572
2.850
0.725
0.505
0.516
2.060
0.303
0.300
0.171
0.878
0.273
0.072
0.384
1.390
0.108
0.077
3.422
2.576
1.049
Particulate Concentration — based on Total of Probe + Cyclone + Filter
   g/m3 (uncorrected)              0.16         0.30         0.13
   g/m3 at12%C02 ,
0.71
0.87
0.51
1.774
0.24
0.58
                                      89

-------
     Analysis for trace  metals in each of the filter and probe plus cyclone fractions was
carried out by spark emission spectroscopy. These results are given in Table 16 for the filter
portion and Table 17 for the probe plus cyclone portion. There is a fair degree of similarity
between the samples of each type, but some large differences in the relative compositions of
the filter and probe plus cyclone catches.
                 Table 16.  TRACE METALS ANALYSIS OF FILTER CATCH
                               (values in ppm of participate)
                                                        Base
       Element


         Al
         Ag
         B
         Ba
         Ca
         Cd
         Cu
         Cr
         Fe
         Ga
         K
         Mg
         Mn
         Mo
         Ni
         Pb
         Sn
         Si
         Sr
         Ti
         V
         Zn
         Na
2
22,000
30
560
5,600
250
140
30
80
10
42,000
800
300
20
3
17,000
1,700
560
30
100
5
25,000
42,000
3
39,000
50
3,500
7,800
400
200
80
200
20
58,000
1,500
400
60
10
23,000
2,700
1,200
80
200
10
35,000
58,000
4
	
40
2,300
9,100
200
200
20
100
10
11,000
2,300
500
—
30
23,000
5,500
3,600
30
—
—
82,000
11,000
5
52,000
30
2,300
2,900
1,700
170
10
100
10
20,000
3,500
900
10
20
20,000
4,000
2,900
20
—
—
69,000
8,600
                                         90

-------
Table 17.   TRACE METALS ANALYSIS OF PROBE AND CYCLONE CATCH
                   (values in ppm of paniculate)
    Element
                               Base
At
Ag
B
Ba
Ca
Cd
Cu
Cr
Fe
K
Mg
Mn
Na
Ni
P
Pb
Sb
Sn
Si
Sr
Ti
Zn
100,000
30
30
1,000
100,000
30
300
300
100,000
3,000
10,000
1,000
3,000
100
3,000
1,000
—
100
30,000
30
3,000
10,000
100,000
100
30
1,000
100,000
100
1,000
300
100,000
10,000
10,000
1,000
3,000
100
3,000
30,000
1,000
1,000
30,000
100
10,000
30,000
100,000
100
30
1,000
100,000
100
3,000
300
100,000
10,000
10,000
1,000
3,000
100
3,000
10,000
300
1,000
30,000
100
10,000
30,000
30,000
100

1,000
30,000
100
300
300
100,000
3,000
3,000
300
3,000
100
1,000
10,000
300
300
10,000
30
10,000
30,000
                             91

-------
                                  SECTION VIM

                              DESIGN GUIDELINES


     Up to now the report has been focussed on the description of the combustion process
using various mathematical models to  delineate the  fluid flow and  patterns within the
furnace, the effect of different types of mixing, and the importance of air distribution and
temperature control. All of these  major topics  can  be  focussed  directly  on the design
guidelines for  incinerators aimed  at reducing combustible emissions  and improving the
combustion efficiency. The important parameters are furnace  temperature, furnace design,
and the air distribution  and flow  rate.  We will  concentrate first on these parameters and
then specifically on the criteria for the design  of an overfire jet mixing system  to take
advantage of the several benefits such a system offers.

TEMPERATURE

     The plot of emission factor as a function of temperature, which was given in Figure 1,
identified two phenomena which could cause significant levels of combustible emissions:

     •   quenching, and
     •   oxygen deficiency resulting from a mixing limitation.

Quenching  became prominent at temperatures below 1200°F, while the limitations in high
temperature mixing were observed at temperatures above  1800°F. There was apparently no
difference in the combustible emission  levels in the range from 1400°F to 1800°F. As a
basic operating guideline, the breeching should therefore be  maintained within this tem-
perature range at all times.

     „,                           .   ,        Table 18, EXCESS AIR  REQUIREMENTS FOR
     These  temperatures  correspond   to  a          EFFICIENT COMBUSTION
range in CO2 concentration of approximately
6 to 10% and  a range in excess air between                           Recommended
100 and 200%. The latter figure can be placed       Fuel                 Excess Air (%)
in perspective  by  considering the recom-       Natural Gas                5-  10
mended percent excess  air requirements for       Fuel Oil                  10-  20
efficient  combustion of  other fuels as shown       Coal                     30'  60
in Table 22.                                     Refuse                 10°-200

FURNACE DESIGN

     The strategy  for combustible  emissions is dependent upon maintaining control of the
various operating  parameters, relying on design  parameters such as furnace configuration
and  relative dimensions. The designer  can considerably  enhance  the  control and mixing
                                         93

-------
      required to reduce emissions. The furnace should be air-tight to avoid in-leakage that can
      quench the combustion or alternatively avoid smoke escaping from the furnace chamber. In
      sealing the furnace, however, care should be  exercised to ensure that the grates  are  still
      readily accessible for maintenance.

      Configuration

          One of the important functions of the  furnace configuration with respect to eliminat-
      ing combustible emissions is to ensure that  the gases are uniformly mixed throughout the
      furnace to minimize  stratification such as that observed at Newton. In our tests, the data
      clearly showed that the hot  gases from the burning zones formed a stable layer in the upper
      portion of the breech, while  the colder gases from the burnout zone stayed  at the bottom of
      the breech. This was evidenced by the  existence of temperature profiles within the breech
      and the existence of high H2 O concentrations in the colder gases caused by water evapora-
      tion in the  ash quench hopper.  The stratification is indicative of the existence of large
      temperature gradients  in the furnace and suggests that zones within the furnace exist where
      quenching does occur. The test data clearly showed that there was a quench zone over the
      burnout grate and indeed high levels of combustible emissions were measured. To avoid this,
      the configuration  of the furnace should be such as to induce front to back mixing, forcing
      the  hotter gases  to pass through the  colder gases  or vice versa.  Several designers have
      suggested furnace  configurations specifically aimed toward these ends.

          Two  configuration attempting  to  achieve this type of mixing are  shown below in
      comparison with the Newton configuration.
Hot gases forced through cold gases
      Height
Cold gases forced through
hot gases
Hot and cold gases remain
stratified
          The height  of a furnace should be as close to the refuse surface as possible without
     effecting flame impingement on the roof. There are three reasons for this constraint:

          1.   The buoyancy of the flame causes the gases to accelerate at a rate which is
               dependent upon the height. The lower roof will effect a lower gas velocity,
               permitting the overfire  air stream to influence  the  overall mixing patterns
               more  strongly.  This  would  result in  better mixing  control  rather than
                                               94

-------
         allowing gases from the bed to dominate the flow patterns in the furnace
         completely.

     2.   The lower velocity makes it easier to achieve a full jet penetration of the
         furnace without introducing so much air that the flame would be quenched.
         This is important in adding flexibility to the jet design.

     3.   The lower gas velocities result in smaller negative pressures at the base of the
         flame and therefore minimize leakage.

     Furthermore, we believe that the height of the roof above the refuse bed should remain
approximately constant throughout the length of the furnace. On many furnaces the roof is
at a constant height, but each grate is successively  lower than the one preceding it  so that
there is a great difference in the height  above the grate at the front and the back of the
furnace. This results in the possibility of  establishing pressure gradients within the furnace.
We observed in-leakage over the third  grate at times when there was out-leakage over the
first grate which confirms this concern.

Length

     The length of the furnace is one of  the primary parameters used in setting the refuse
residence time. To achieve a  good burnout, we believe that residence times on the order of 1
hour are required. Shorter residence times could be used if burning rates higher than 60
lb/hr/ft2 were attained.  In many  configurations the length of  the furnace also is the
predominant factor in  determining the  gas  residence time. The required value  for the
gaseous residence  time is  discussed  below in conjunction with the jet design; times on the
order of 1  second are not unreasonable  in large incinerators. Smaller units having much
higher mixing intensities have demonstrated effective residence times as low as 0.5  second.

Width

     The width of the furnace is of primary  importance  in specifying jet  penetration
discussed below, but  it is more often set to attain a desired feed rate for a given  furnace
length and burning rate.

UNDERFIRE AIR DISTRIBUTION AND FLOW RATE

     The primary air flow in the furnace is the underfire air.  The capacity of the underfire
air system should be at least 150% of stoichiometric air requirements with the flexibility of
being able  to deliver  100%  of stoichiometric requirements under any given section of the
grate, especially in the sections of the grate where most of the burning occurs.

     Each section  of the  furnace should  have two important design features. First, the air
plenum should have the capability  to subdivide and individually control  the air rates and,
                                           95

-------
second, the grates  should be designed to allow an even distribution of air over the entire
surface of the grate. The latter requires that the wind box be sealed so that high undergrate
pressures  can be attained, and that the  grate have  a high pressure  drop with respect to
refuse.  In Newton, the grates  have a relatively  low pressure drop  and the non-uniform
distribution of underfire air was clearly seen. In particular, we observed the channelling of
air around rather than through the bed.

     The  theoretical air requirements for  various  fuels, including  refuse, are given in
Table 23.  As a rule of thumb, 10 scf of air are required per  1000 Btu's of heat liberated. The
total air requirement is calculated from the equation :
                 Q .           (HHV, « s,cic, (TPD) -                           (50,
                      1 000 Btu                       1 440 mm/day
where

                   Q     =    fan output
                   HHV  =    refuse heating value (approximately 4500 Btu/lb).

For example, a 250-TPD furnace requires 23,400 scfm at  150%  of stoichiometry (50%
excess  air). The air handling capacity of the  grate is set  by the burning rate rather than the
feed rate:
         Table 19. THEORETICAL AIR REQUIREMENTS OF REFUSE AND OTHER FUELS

               Fuel                Ib Air/lb Fuel       Ib Air/MM Btu       scf Air/M Btu
     Refuse                           3.22              724                 9.77
     Wood                            3.29              703                 9.49
     Peat                             2.33              722                 9.75
     Lignite                           5.27              746                10.07
     Sub-Bituminous B                  7.58              740                10.00
     Bituminous (High Volatile)           9.08              742                10.01
     Bituminous (Volatile)               10.99              760                10.26
     Anthracite                        9.23              831                11.22
     Fuel Oil                          13.69              750                10.13
     Methane                         17.26              722                 9.75
                                          96

-------
     For a burning rate of 60 lb/hr-ft2, the underfire rate is 68 scfm/ft2  or about 8000
scfm under  a 8- by 15-foot grate. Burning rates much higher than this are possible and the
designers must ensure that the grate is not limited by gas handling capacity. A safety factor
of 5 applied to the burning rate, i.e., assuming a rate of 3000 lb/hr-ft2 , is not unreasonable
in the absence of actual performance data for a particular design. The impact of this design
strategy is felt in these areas: burning rates, temperature control, and mixing.

Burning Rates

     Our test data showed that the burning rate tended to remain constant at approximately
60 lb/hr-ft2 , the reason being that the bed  requires a minimum amount of oxygen to burn
and  naturally will  draw in air from overfire  to  make up  oxygen not  supplied with the
underfire air. As the underfire air rate increased, the overfire air entrainment decreased. We
concluded that the 60 lb/hr-ft2 was a minimum burning rate resulting from the fire's natural
tendency to  draw in oxygen  to sustain combustion. The relatively constant burning rate
exhibited for less than stoichiometric air flows confirms the existing  design rule for setting
burning rates at 60 lb/hr-ft2, but it does not suggest that this burning rate is fixed regardless
of operating conditions. With positive control and adequate underfire air fan capacity,
reliance on the overfire air is minimized  and the burning  rate is  no  longer  limited by
naturally available  overfire oxygen. As a result  higher burning rates are attainable. Not
enough data are available to  predict limiting burning rates, but values as high as 300
lb/hr-ft2 have been reported.

Temperature Control

     Underfire flow rates  greater than  the stoichiometric requirement also enhance the
capability  to  control furnace temperature, particularly within the bed and at the grate.
Consider the schematic diagram shown  below representing the refuse bed temperature as a
function of stoichiometric air flow rate. At air flows less than the stoichiometric require-
ments  all  of  the oxygen  will be consumed,  resulting in the generation of combustion
products at the adiabatic flame temperature. The  gases leave the bed at  approximately the
              Refuse
               Bed
              Temp
                                         100%

                                   Percent Stoichiometry
                                           97

-------
flame temperature — less any cooling of the gases — as they pass through the refuse bed. The
occurrence of overfire entrainment compounds the problem by tending to stabilize the heat
release rate, regardless of the underfire air setting.

     The small slope of the curve above indicates that low underfire air rates provide very little
refuse  bed  temperature  control. On  the  other  hand,  for  flow rates  greater than the
stoichiometric requirement, the additional air tends to dilute the combustion products so
that an increase in  the air flow results in a fairly rapid decrease in the bed temperature. This
type of temperature  control  would be of  considerable importance to the incinerator
operator, but the concept is not currently in practice except in the newest units which
provide adequate underfire air. Care must  be exercised, however, since a large amount of
excess  air promotes quenching and particulate entrainment, particularly  over the burnout
section  of the  grate.  The latter concern is minimized because most incinerators will soon
have high  efficiency  air pollution  control devices, and also because oxygen provided by
overfire air currents tends to entrain particulate.

Mixing

     Control of the underfire flow rate permits much better temperature control over the
bed and effectively eliminates large variations in temperature. As a result quenching zones
are eliminated and stratification is minimized.

OVERFIRE AIR JETS

     The advantages of using overfire air jets are temperature control, increased mixing, and
possibly slag reduction. During the course of this work it was very apparent that the sidewall
jets were much more  effective  in controlling temperatures in the 1400°F  to  1800°F range
than were the existing roof jets, the reason being that jets, when properly designed, attain a
full penetration of  the furnace gases and result in a much more uniform mixing than do the
roof jets. In addition, the jets can be  effective in inducing a uniform side-to-side mixing,
which is easily  observed as well as increasing the mixing intensity. However, because of the
furnace dimensions we believe that front-to-back  mixing  will  be more readily achieved by
proper  furnace  design. During the tests we  were not able to demonstrate a  significant
increase in mixing  intensity; however,  from these tests we have derived criteria for evalu-
ating the  types of  furnace conditions under which  overfire air jets could be effective in
inducing the fine-scale turbulence required to improve the combustible emissions burnout.

Additional Uses for Jets

     Finally, others have suggested that jets can also be effective for reducing slag formation
on  incinerator walls. We were unable  to  confirm  this  observation.  Certainly the  early
tempering of the gases with side wall jets rather than roof jets favors reduced slag formation,
but the increased momentum of the particles induced by the high jet velocities can easily
                                          98

-------
increase the collection efficiency of slag on impact with the wall. Hence, care should be
exercised  in  placing the jets to avoid inducing impingement of entrained particulate on
furnace walls.

     We did  observe an increased rate of slag buildup directly over the jet opening and
recommend the jets be surrounded by a slag-resistant refractory, such as silicon carbide.

Design Criteria

     The two parameters of importance in designing overfire air jets are the jet penetration
and  the position of the jets.  The method of Ivanov for determining jet penetration was
verified as a result of the testing program. With respect to the jet configuration or spacing,
several variations were tried including both opposed jets, interlaced jets, and jets acting only
on one side of the furnace; all were equally effective. An additional point, not considered in
earlier studies is that the jets should be located as close as possible to the burning surface for
the following reasons:

     a.    During the tests one could clearly observe a reduction in the flame height as
         a result of the overfire air jets. In furnaces designed with relatively low roof
         heights, as we have  recommended, this aspect is very important for limiting
         the amount of flame impingement on the roof, thereby avoiding potential
         refractory damage.

     b.   The temperature control and uniform mixing are more easily achieved when
          the gas velocity off the refuse bed is relatively low. This permits a greater
          control of the macroscale mixing and also results  in the attainment of a full
         penetration with  a minimum  of overfire air so that quenching  can be
         avoided.

     The procedure for designing jets is outlined as follows:

     a.    Determine overfire  air requirement based upon temperature control -

          Example:
          Furnace size - 250 TPD
          Stoichiometric air — 15,600 scfm
          Underfire air (approximately 150%) - 23,400 scfm
          Overfire air (approximately 100%) - 15,600 scfm

     b.   Set jet size and number -

          In practice, the jet diameter is limited in size only  by the tendency of the jet
          to slag over. A secondary effect on jet power output, as discussed below, has
                                         99

-------
     a significant effect only  on small-diameter jets, but the design is usually
     constrained by the slagging problem. A 3-inch  jet diameter is  about the
     smallest practicable. Jet spacing should be between 3 and 5 jet diameters.

     Example:
     Furnace length — 40 feet
     Jet diameter — 4 inches
     Spacing — 16 inches
     Number = 40 feet/1.33 feet/jet = 30 jets

c.   Set jet penetration and configuration —

     The various possible configurations — jets on one side, jets opposed on both
     sides, jet interlaced on both sides — appeared to be equally effective. Ivanov
     has suggested that jets on only one side have a slight advantage. The selection
     of a configuration will depend upon considerations such as slag reduction.

     Example:
     Configuration — one side
     Penetration (furnace width) - 8 feet

d.   Determine jet velocity

     Since the total overfire air and the number of jets are specified above, the jet
     velocity can be calculated. Using Eq. (6), the required crossflow velocity can
     be calculated as a function of penetration:


     ui = Qr/N|Lj J  _       15,60Qscfm      =  100 ft/sec.
               V4 /      7r(30)(60)/l/3fP\                                  (52)
e.    Calculate required crossflow velocity (Eq. (4)):


                                                        2260 R\_ 11 -7 ni—„    (53)
            ^
                                    100

-------
     f.    Estimate jet height above refuse (Eq. [9]):
          Temperature ratio* - 2260°R/540°R = 4.2

                       23,400
       uc = Q/Ag = (40> x 8')60   =   l -2 ft/sec-
                          1ii   _  i                                              (54)
                          TC      J
              L
= [(13.7)2 -(1.2)2 ft2/sec21/2.32 ft2/sec2 (3.2) = 0.9 ft
     The outline above illustrates only one method for arriving at a consistent jet design.
Based upon an initial air flow criterion, the order of the calculations could be changed to
base the design procedure on a mixing criterion, considering the primary jet parameter to be
jet penetration rather than total overfire air rate. An example in which this design approach
would be more useful would be designing the mixing system for controlled-air incinerators
where  the  underfire air is severely  limited so that the introduction of large amounts of
overfire air does not result in quenching.

Mixing Intensity Criteria

     One of the methods for reducing combustible emissions is through an increase in the
intensity of turbulence  in the overfire region. The power required to attain a given level of
mixing depends upon the relative densities of the fluids being mixed, but more importantly
upon the size of the furnace. As a general rule, the required mixing power scales with the
fifth  power of  the  characteristic  dimension of  the  system.  Equation (24) gave the
relationship for 0 in terms of air flow rates.

     The total jet flow rate is determined as a product of the individual jet flow rate
(Eq. (44)) times the grate area. In most cases the grate area  is  simply  the product Lp • Lf.
Substituting in these values the total  flow rate QT : is determined from Equation (55):
                                                                                 (55)

The above relationship is  dependent upon such things as the spacing parameter and  the
furnace dimensions, but is not dependent upon the jet diameter or jet velocity. Hence, in a
properly designed jet system the total air flow rate introduced via overfire air jets depends
 *Note: In a sealed furnace the temperature ratio becomes 2260°R/1460°R or 1 .55 so that:

             u = 8 ft/sec.
              U
             Az = [ (82 - 1 .22 )ft2 /sec2 ] /2 (32 ft/sec2 )0.55 = 1 .8 ft
                                          101

-------
upon the furnace and combustion parameters, but not on the jet design itself; i.e., the total
flow rate of the jets is fixed for a given application. The power output of the mixing jets,
however, is a function of the jet design, specifically being inversely proportional to the cube
root of the jet diameter. For example, if a jet system is designed using 1-inch jets rather than
4-inch jets, the power output is increased by a factor of approximately 1.6. In most cases,
the jet design will be constrained more by the occurrence of slagging preventing the use of
small-diameter jets than  by the microscale power input considerations.  For this reason and
others the microscale considerations do not play an important role in the jet design itself.
However, the microscale relations can be of considerable importance in determining which
furnace designs  could benefit from the use of overfire air jets for microscale  mixing and
which furnaces would benefit from overfire jets only on the  macroscale; i.e., for increased
temperature control, slag  reduction, and  reduced  flame heights. To be effective  on the
microscale  the factor /3t should have a value of approximately 1/2.  The procedure for
determining the potential effectiveness of overfire jets is as follows:

     a.   Determination of j301

         The initial value for  the mixing intensity is dependent  upon the furnace
         configuration  and dimensions based upon an application of  Eq. (36). The
         example considered below illustrates the use of these equations.

         Example:
         Consider the furnace jet system in the example given under design criterion
         above:
                        2/3
                                                         (56)
            P0t = 0.15
                       (2700 ft3/sec) (1  sec)
(320 ft2) 8 ft
                                            2/3
                     = 0.19
                                                         (57)
     b.  Determine mixing intensity with jets:
                                  320 ft2 \   / 270 ft3 /sec
                                   2.6ft2/   \2700 ft3 /sec
                                )'I
                                                                                  (58)
       In this case, the incinerator would benefit from an installation of jets.
                                          102

-------
                                  REFERENCES

1.    Systems Study of Air Pollution from Municipal Incinerators, Arthur D.  Little, Inc.,
     Report to EPA, Contract CPA-22-69-23, March 1970.

2.    Incinerator Overfire Mixing Study, Arthur D. Little, Inc., Report to EPA, Contract
     EHSD 71-6, February 1972.

3.    Ivanov, Yu V., Effective Combustion of Overfire Fuel Gases in Furnaces (translation),
     Published by the Estonian State Publishing House, Tallinn, Estonia, U.S.S.R., 1959.

4.    Hawthorne, W.  R.,  Weddell,  D.S.,  and Hottel,  H.C., Mixing and Combustion in
     Turbulent Gas Jets, Proc. of Third Symposium on Combustion and Flame Explosion
     Phenomena, the Wilkins and Wilkins Company, Baltimore, Md., 1949.

5.    Corrsin, S., Simple Theory of an Idealized Turbulent Mixer, Jour. A.I.Ch.E., Vol. 3,
     No. 329, 1957.

6.    Turbulent  Flows and Heat Transfer, Volume V:  High-Speed  Aerodynamics and Jet
     Propulsion, C.C. Lin, Ed., Princeton Univ. Press, Princeton, N.J., 1959.

7.    Turbulence — Classic Papers on Statistical Theory, Friedlander, S.K., and Topper, L.,
     Eds., Interscience Publishers, Inc., New York, 1961.

8.    Comprehensive Test Program — Particulate Emissions and Baffle Washer Performance
     at the Lower Merion, Pennsylvania, Incinerator and the Newton, Massachusetts, Incin-
     erator, Report for the Research Corporation of New England, Metcalf and Eddy, Inc.,
     Boston, Mass., 1970.
                                       102a

-------
                                    APPENDIX A

                         TEST MEASUREMENT AVERAGES

     Within this appendix, the averages of all the test measurements made during the testing
program have been tabulated. The analytical data are listed in Table A-l. A summary of the
various operating conditions under which each test  was run is given in Table A-2 for the
Series 20 tests, and in Table A-3 for the Series 30 tests. Finally a summary of the material
and energy balance is given in Table A-4.

     In addition, we  have included actual test data from the initial four-week series of tests
during which time the analytical procedures were being verified. These data are summarized
in Table A-5. During these tests we noted severe probe leakage problems and difficulties in
measuring  velocity data.  Therefore, the absolute quantities listed in the table are often
incorrect,  as can be noted by comparing the reported value of CO2 (corrected for moisture)
with the data given in Figure 26. However, the relative values  should still be  accurate and
can be used as additional data for verifying such things as the existence of the water gas shift
equilibria.  Abnormally low velocities (less  than ~ 35ft/sec) are a result of velocity probe
plugging. Data in Figure 28 can be used to obtain converted readings. Finally, the measured
H2O was  usually lower than the actual moisture  content during the first weeks of testing
before the new analytical technique for H2O was used.  Data in Figure 25 can  be used to
obtain corrected readings for the value. We have not relied on these data for confirming the
various models presented, but  we have included them in this report because much of the
data is valid; considering the tremendous lack of actual test data on a municipal incinerator,
even inconsistent data  (when recognized as such) convey more information than has been
available previously.
                                         103

-------
                                                                                                                    Table A-1. ANALYTICAL DATA
DATA
BREECH ZONE 1

Run
BL
JET
BL
BL
JET
BL
JET
BL
JET
JET
BL
BL
JET
JET
BL


- 22/1
- 22/1
- 22/2
- 23/1
- 23/1
- 23/2
- 23/2
- 24/1
- 24/1
- 24/2
- 24/2
- 25/1
- 25/1
- 25/2
- 25/2

CO2
10.77
8.73
10.43
9.76
7.25
8.90
5.70
6.61
9.39
8.71
10.03
11.80
8.35
7.74
8.45
% (Wet)
02
6.07
8.95
7.08
8.36
11.02
7.85
11.33
10.44
7.73
9.14
6.53
5.40
7.11
11.80
10.78
PPM (Wet)
H2O
14.50
12.20
15.20
13.80
10.04
16.20
9.50
14.40
14.00
10.80
16.20
12.90
14.30
7.80
11.90
CH4
5.13
3.29
11.87
12.28
1.19
1.46
0.90
2.56
3.44
0.89
1.25
22.21
26.99
1.38
1.05
CO
0.85
0.21
0.84
103.44
52.64
245.11
119.91
79.18
47.30
40.14
341.48
641.27
679.17
64.54
64.31
CO 2
5.38
5.56
7.00
5.46
4.74
3.84
4.75
6.11
6.09
5.79
5.77
8.15
6.74
5.81
3.99
BREECH ZONE 2
02
13.54
11.83
13.00
14.36
14.69
14.32
13.14
11.83
11.46
12.19
11.93
11.11
13.34
11.19
17.14
H2O
7.10
6.80
7.00
7.30
5.50
7.00
5.80
2.40
9.50
7.30
8.40
11.40
9.50
5.90
4.90
CH4
9.29
12.11
12.00
10.66
9.92
10.23
16.95
12.36
15.15
9.73
15.57
14.39
8.37
5.17
7.84
CO
455.21
270.28
142.50
308.22
359.10
1,185.75
1,071.'52
272.51
9.05
11.58
11.45
338.89
259.43
275.24
461.23
CO2
0.71
2.21
1.54
0.25
0.12
0.60
0.77
0.26
1.83
1.06
1.44
2.50
3.28
1.79
1.17

i cmrcnMi unc i ri
BREECH ZONE 3
02
18.82
18.03
18.72
19.65
20.06
19.94
19.19
18.55
16.35
17.38
16.94
17.88
17.76
19.33
19.95
H2O
3.10
3.80
3.40
4.10
2.10
2.70
2.80
4.50
5.10
3.40
3.70
5.50
5.40
3.20
2.20
CH4
4.84
6.73
14.94
2.87
3.18
10.37
20.16
6.44
23.96
21.01
35.39
30.71
42.57
24.20
22.49
CO
77.76
50.50
119.52
9.59
9.79
608.12
510.30
23.87
249.11
217.35
300.93
571.72
339.37
307.34
501.22

Zone
1
1,721
1,508
1,700
1,715
1,437
1,717
1,421
1,647
1,694
1,551
1,730
1,900
1,796
1,400
1,523

II
1,446
1,387
1,458
1,446
1,125
1,350
1,158
1,385
1,470
1,341
1,426
1,696
1,587
1,246
1,131

III
1,029
1,004
1,100
1,046
737
871
856
892
1,152
949
1,071
1,283
1,225
875
742

Avg.
1,366
1,272
1,411
1,372
1,058
1,265
1,123
1,271
1,424
1,258
1,382
1,598
1,505
1,143
1,097
BREECH FLOW
Velocity (ft/sec)
1
51.80
50.41
54.17*
53.02*
44.51*
52.12*
48.19*
55.74
56.50
51.42
53.93
54.94*
51.38
47.50*
49.31
II
48.28
46.58
47.66
51.74
44.54
49.05
49.01
47.54
50.84
44.52
47.15
54.86
52.53
44.45*
44.41
III
42.78
44.77
38.87
44.23
37.08
40.81
39.81
42.26
45.59
40.23
44.08
48.24
48.57
42.22
36.30
Total
Ib. moles/hr
7,643
8,007
7,376
7,935
8,102
8,026
8,439
8,210
7,929
7,742
7,705
7,496
7,587
8,186
8,142
BL
JET
JET
BL
JET
JET
BL
JET
JET
BL
JET
JET
JET
- 31/1
- 31/1
- 31/2
- 32/1
- 32/1
- 32/2
- 33/1
- 33/1
- 33/2
- 34/1
- 34/1
- 34/2
- 34/3
11.11
7.96
8.22
9.27
9.60
8.05
10.54
8.82
9.08
9.30
8.63
8.00
9.20
6.16
9.66
9.28
6.82
6.98
8.05
6.25
8.38
6.45
7.02
8.37
9.82
7.17
13.20
10.50
11.60
14.65
12.70
13.40
14.30
12.65
15.10
12.20
12.75
9.00
8.00
40.10
2.50
0.88
9.96
2.18
4.15
21.00
8.03
267.43
46.09
6.63
3.18
3.68
145.56
30.43
17.68
187.77
165.87
157.61
241.67
63.32
382.05
210.72
47.98
40.95
41.40
7.33
5.09
5.85
7.76
7.91
6.59
6.56
6.76
6.93
6.88
5.85
5.04
6.50
10.49
13.15
12.63
10.56
8.89
9.98
12.02
11.26
11.04
10.92
11.53
12.62
11.33
9.50
7.35
7.10
9.70
11.10
10.25
7.50
9.85
8.70
8.20*
8.45
6.50
7.10
28.96
4.16
25.54
6.32
5.33
9.87
10.82
8.11
219.12
5.69
11.62
9.81
12.07
162.90
187.61
278.70
31.60
120.01
229.49
300.62
234.39
314.98
183.60
244.89
250.11
222.96
2.65
1.25
1.61
2.26
3.30
2.31
2.06
3.09
2.98
1.63
1.41
1.90
1.34
16.38
18.10
17.99
14.90
13.78
15.33
17^00
16.21
16.12
16.09
15.97
16.40
16.86
5.30
3.20
4.80
5.65
8.10
7.60
6.05
6.25
6.80
3.60
5.50
4.60
4.20
38.44
5.42
4.76
32.07
34.46
62.64
42.93
76.87
950.64
17.83
28.06
73.93
21.07
181.82
20.32
19.04
264.18
22.97
200.60
420.42
517.96
829.48
337.40
304.76
453.15
411.94
1,750
1,507
1,525
1,825
1,683
1,698
1,777
1,666
1,753
1,793
1,554
1,497
1,597
1,543
1,261
1,308
1,707
1,667
1,609
1,545
1,485
1,555
1,478
1,417
1,367
1,481
1,193
836
1,025
1,381
1,408
1,344
1,251
1,240
1,306
1,150
1,106
1,023
1,115
,477
,179
,267
,621
,576
,538
,503
,446
,512
,458
,354
,277
,398
53.26
52.28
44.20
54.55*
51.33
51.10
48.02
45.77
45.45*
53.93*
49.85*
48.85*
49.93
48.91
47.91
42.85
51.22
52.42
50.28
46.25
43.73
43.00
47.43
42.43
44.86
48.65
43.93*
38.24
38.25
48.64
48.76
46.95
42.79
42.78
46.69*
41.61
37.19
41.10
35.52
7,369
8,231
7,086
7,258
7,324
7,260
6,824
6,783
6,723
7,280
6,984
7,584
7,023
'Approximate-based on estimates due to plugging of velocity measurement probe during a run.
                                                                                                                             104

-------
                                                                                                           Table A-2. SUMMARY OF VARIOUS OPERATING CONDITIONS - SERIES 20 TESTS
                                                                       Operating Parameters
                                                                                                                                                                         Analysis of Operation

BL
Jet
BL
BL
Jet
BL
Jet
BL
Jet
Jet
BL
BL
Jet
Jet
BL
Purpose
22/1 Exploratory runs
22/1 Exploratory runs
22/2 Exploratory runs
23/1 Exploratory runs
23/1
23/2
23/2
24/1 Exploratory runs
24/1
24/2
24/2
25/1 Exploratory runs
25/1
25/2
25/2
Procedure
No sidewall or roof jets
Sidewall jets on — alternative
opposed; 50% Penetration
No sidewall or roof jets — same
as BL22/1
No roof or sidewall jets
Alternate, opposed sidewall jets
on-penetration ^60%
Same as BL 23/1 except smaller
UFair
Same jet condition as Jet 23/1
except more overf ire and less
UFair
No sidewall or roof
jets
Alternate opposed jets, lower
UF air from Bl 24/1
Same jet settings as Jet 24/1
Same as Jet 24/2 except no
jets on
No jets sidewall or roof
Alternate opposed sidewall jets
or penetration "u 25%
Stone feed rate and raise
to the jet and UF in Penetration
40%
Same as Jet 25/2 except no jets
Under Fire Air
CFM | (%) || (%) in (%(
8400 11 30 59
7400 10 32 58
7400 10 32 58
9800 18 31 51
9600 17 31 52
9300 17 32 51
9100 16 35 49
8400 14 37 49
7000 11 39 50
7250 10 41 49
7250 10 41 49
8450 9 43 48
7450 10 43 47
8250 9 41 50
8250 9 41 50
Over Fire
Air
(CFM)
0
4900
SJ
0
0
4300
SJ
0
5700
SJ
0
2800
SJ
2800
SJ
0
0
2200
RJ
3600
SJ
0
Refuse
Wet/Dry Feed Rate Grate Speed (FPH)
(TPH) 1 II III
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
9.8 45 45 45
9.8 45 45 45
9.8 45 45 45
8.74 40 45 50
8.74 40 45 50
8.74 40 45 50
8.74 40 45 50
9.78 45 56 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
8.74 40 45 50
8.74 40 45 50
Average
Res. Time Breech Temp.

-------
                                                                                                       Table A-3. SUMMARY OF VARIOUS OPERATING CONDITIONS - SERIES 30 TESTS
 BL   31/1
Jet   31/1
Jet   31/2
    Purpose

Determine effect of
blocks of jets over
drying grate & 2nd
   grate
           Procedure

No sidewall or roof jets
run normal conditions
Demonstrate effect of  Jets 1-5 on both sides
sidewall jets over the   penetration 'V 50%
drying grate and drop
zone to the second
grate

Demonstrate effect of  Jets 6-10 or, both sides
sidewall jets over the   penetration 'Vi 50%
2nd grate
 BL   32/1     Determine effect of   Open roof jets, half open
               placement of jets on
               roof as opposed to wall
               and effect of reducing
               feed, U Fair, and SJ
               air by same proposition
               Demonstrate effect of
               roof jets

 Jet   32/1     Demonstrate effect of Use sidewall jets —
               sidewall jets          alternate opposed — both sides

 Jet   32/2     Demonstrate effect of Reduce feed grate speed by 10%,
               sidewall jets with      reduce SJ air by 10%,
               lower flow, lower      reduce UF air by 10%
               UFair, and lower
               feed rate
CFM

6400
                                                                       6400
                                                                                                Operating Conditions
                                                                                     Under Fire Air                           Refuse
                                                                                                              Jets                 Feed Rate
                                                                                I (%)     II (%)     III (%)      CFM     Wet/Dry       (TPH)
                                                                                 11
                                                                                 11
                                                        6750      22
                                                                                          39
                                                                                          39
                                                        6800      23       31
                                                                                                   50
                                                                                                   50
                                                                                    46
                                                                           31       47
                                                                                                             5535
                                                                                                              SJ
                                                                                                                        Dry
                                                                                     Dry
                                                                       6400      11       39       50        4750      Dry
                                                                                                              SJ
                                                                                               1200       Dry
                                                                                                RJ
                                                        7200      22       29       49         4195       Dry
                                                                                                SJ
                                                                                               3790       Dry
                                                                                                SJ
                                                                                                                                     8.74
                                                                                                                       9.78
                                                                                                                       8.74
                                                                                                                                   Grate Speed (FPH)
 I

40
II    III

45    45
                                                                                                             40     '  45    45
                                                                                                                                     8.74
                                                                                                  8.74        40       45    45
                                                                                                                       9.78        45       45    40
                                                                                                                                  45       45    45
                                                                                                                                  40       40    45
Res. Time
   (hr)

   .96
                                                                                                                                        .96
                                                                                                                                        .96
                                                                                                                                                             .98
                                                                                                                                                             .93
                                                                                                                                                            1.00
                                                                                                                                                      Average
                                                                                                                                                    Breech Temp.
                                                                                                                                                                                           1460
                                                                                                                                                                                           1179
                                                                                                                                                       1267
                                                                                                                                                                            1619
                                                                                                                                                                            1576
                                                                                                                                                                            1538
                                                                                                                                     Analysis of Operation
                                                                                                                                  High Furnace      Low Furnace
                                                                                                                                  Temperature       Temperature
                                                                                                                                                                                                              1850
                                                                                                                                                                                                             1620
                                                                                                                                                                          1560
                                                                                                                                                                                               1890
                                                                                                                                                                                               1750
                                                                                                                                                                                               1840
                                                                                                                                                                                                                               1400
                                                                                                                                                                                                                               1290
                                                                                                                                                                                                                               1385
                                                                                                                                                                                                                               1660
                                                                                                                                                                                                                 1670
                                                                                                                                                                                                                 1560
                                                                                                       Average Stack
                                                                                            Relative     Temperature
                                                                                            Variation        ( F)
                                                                                                                                                                                                                                            large
                                                                                                                                                                      large
                                                                                                                                                                                                                                            med
                                                                                                                                                                                                                                             med
                                                                                                                                                                                                                              small
                                                                                                                                                                                                                                                             320
                                                                                                                                                                                                                                                             330
                                                                                                                                                                                                                                                             320
                                                                                                                                                                                                                                                             340
                                                                                                                                                                                                                                              340
                                                                                                                                                                                                                              med-large        360
  Burnout
 Efficiency      Run Times
(106Btu/hr)        (hr)

   98.2        1105-.-1215
                                                                                                                                                                                                                                                                             80.2        1425*1535
                                                                                                                                                                                                                                                                             75.1        1645*1715
                                                                                                                                                                                                                                                                             100.5        1040*1150
                                                                                                                                                                                                                                                              100.9
                                                                                                                                                                                                                                                               96.8
                                                                                                                                                                                                                                                                           1345*1415
                                                                                                                                                                                                                                                                           1545*1655
BL   33/1
Jet   33/1
Jet    33/2
BL    34/1
Jet    34/1
Jet   34/2
Jet    34/3
Determine effect of
one sided vs opposed jets

Determine the effects
of jet placement —
all on one wall or  placed
on both walls

Demonstrate effect
of all jets on one
wall

Demonstrate effect
of jet on both
wallj
                                        Baseline measurement
                                        Use no side or roof
                                        jets
                                        Use all jets on North wall
                                        of incinerator
                                        Use alternate opposed
                                        jets configuration
               Determine the effect
               of jet penetration of
               the operation of the
               furnace
                     Baseline measurement
                     no side jets, no roof jets.
Demonstrate effect of  alternate opposed jets on
partial penetration     all jets set to 25% penetration

Demonstrate effect of  alternate opposed jets on all
total non-overlapping   all jets set at 50% penetration
penetration

Demonstrate effect of  alternate opposed jets on
maximum possible     jets over 2nd grate on maximum,
penetration           jets over drying grate & fall atea
                     on about 3/4 (50% penetration)

ND-No Data
                                   7000    21.43    33.57     45.00
                                                                       7850      20       32       48
                                                                                     Dry
                                   6900    23.19    31.16   45.65        4260       Dry
                                   6900    23.19    31.16    45.65        4260       Dry
                                                                                     Dry
                                                                       7100      20       30       50        4880       Dry
                                                              8.74        40       50    45
                                                                                                                                     8.74        40       50    45
                                                                                                                                     9.78        45       50    45
                                   7100      20       30       50         2466       Dry          8.74         40      50    50


                                   5450      22       31       47         4487       Dry          8.74         40      50    50
                                                                                                                                     8.74        40       50    50
                                                                                                                                                  I 06
                                                                                                                                        .93
                                                                                                                                                                           .93
                                                                                                                                                                           .90
                                                              8.74        40    ,   50    45         0.93
                                                                                                                                                                           0.89
                                                                                                                                                                           0.89
                                                                                                                                                                           0.89
                                                                                                                                                       1503
                                                                                                                                                                                           1446
                                                                                                                                                                                           1522
                                                                                                                                                                                           1455
                                                                                                                                                                                           1354
                                                                                                                                                                                           1277
                                                                                                                                                                                           1398
                                                             1840




                                                             1690



                                                             1910



                                                             1890




                                                             1630


                                                             1010



                                                              ND
                                                                                                                                                                                            1640
                                                                                                                                                                                                                               1550
                                                                                                                                                                                                                               1590
                                                                                                                                                                                                                                1660
                                                                                                                                                                                                                                1470
                                                                                                                                                                                                                                1400
                                                                                                                                                                                                                                ND
                                                                                                                                                                                                                                             small
                                                                                                                                                                                                                                             large
                                                                                                                                                                                                                                             med
                                                                                                                                                                                                                                             small
                                                                                                                                                                                                                                             med
                                                                                                                                                                                                                                             ND
                                                                                                                                                                                                                          300
                                                                                                                                                                                                                                                             300
                                                                                                                                                                                                                                                             320
                                                                                                                                                                                                                                                             320
                                                                                                                                                                                                                                                             320
                                                                                                                                                                                                                                                             310
                                                                                                                                                                                                                                                             ND
                                                                                                                    87.9        1030*1140




                                                                                                                    84.5        1250*1350



                                                                                                                    93.1        1630*1700



                                                                                                                    88.6        1100*1150




                                                                                                                    81.2        1400*1500


                                                                                                                    79.7        1530»1600



                                                                                                                    80.8        1700*1720

-------
                                                                         Table A-4. SUMMARY OF MATERIAL AND ENERGY BALANCE
BL
JET
BL

BL
JET
BL
JET

BL
JET
JET
BL

BL
JET
JET
BL
BL
JET
JET

BL
JET
JET

BL
JET
JET

BL
JET
JET
JET
22/1
22/1
22/2

23/1
23/1
23/2
23/2

24/1
24/1
24/2
24/2

25/1
25/1
25/2
25/2
31/1
31/1
31/2

32/1
32/1
32/2

33/1
33/1
33/2

34/1
34/1
34/2
34/3
CARBON BALANCE
Ib moles/hr
Cln
III
As Refuse
425
425
425
375
375
375
375
425
425
425
425
425
425
375
375
375
375
375
425
425
375
375
375
425
375
375
375
375
Gout
VXU I
AsCO2
400
425
450
375
300
325
300
325
450
375
425
525
450
400
350
500
375
350
450
500
400
425
400
400
425
375
375
400
% Deviation
Cln COut
Cln
4.4
-0.8
-6.8
-1.5
21.1
12.8
19.2
21.9
-6.6
8.3
-0.5
-28.0
-6.9
-5.1
7.0
-34.0
-1.3
5.0
-7.8
-19.0
-6.9
-11.7
-9.6
3.5
-13.0
1.2
1.8
-7.6
                                                       HYDROGEN BALANCE

Hln
III
As Refuse
625
650
700
600
475
500
300
525
700
600
650
850
700
625
550
800
600
550
700
775
625
650
650
625
650
600
575
625
Ib moles/hr
Hln
III
AsH2O
525
550
575
500
375
425
475
425
575
500
550
700
575
500
450
650
500
450
575
650
525
550
525
525
550
475
475
525

Hout
V^U L
AsH20
1,200
1,150
1,250
1,250
900
1,300
950
1,450
1,500
1,050
1,400
1,450
1,400
900
950
1,350
1,100
1,100
1,400
1,550
1,500
1,250
1,250
1,200
1,150
1,250
1,000
900
% Deviation
Hln HOut
Hln
-3.6
3.5
2.8
-16.8
-5.8
-36.6
-12.5
-53.0
-15.5
2.6
- 14.9
6.5
-9.7
22.1
3.0
7.2
-2.8
-5.7
-9.4
-7.6
-30.0
-2.9
-7.7
-4.7
6.0
-17.0
5.3
21.8
OXYGEN BALANCE
Ib moles/hr
o
As Refuse
j;
250 "
260
280
240
180
200
190
200
280
240
260
330
280
240
220
310
240
220
280
310
250
260
250
250
260
230
230
250
o
AsH2O

260
270
290
250
190
210
200
210
290
250
270
350
290
260
230
330
250
230
290
320
260
270
270
260
270
240
240
260
0
As02

3,000
3,200
2,900
3,100
3,200
3,300
3,400
3,200
3,100
3,100
3,000
2,800
2,900
3,300
3,200
2,900
3,300
2,800
2,800
2,800
2,800
2,600
2,600
2,600
2,900
2,700
3,000
2,800
°0ut
vy u i.
AsC02

800
850
900
750
600
650
600
650
900
750
800
1,050
900
800
700
1,000
750
700
900
1,000
800
850
800
800
850
750
750
800
°0ut
VSUl
AsO2

2,000
2,100
1,800
2,300
2,600
2,400
2,500
2,300
1,900
2,000
1,900
1,800
2,000
2,400
2,700
1,800
2,300
1,900
1,600
1,500
1,700
1,700
1,700
1,600
1,700
1,700
2,000
1,700
°0ut
VSUv
AsH2O

600
575
625
625
450
650
475
425
475
525
700
725
700
450
475
675
550
550
700
775
750
625
625
600
575
625
500
450
% Deviation
°ln °0ut
°ln

2.1
3.2
2.3
-4.6
.8
2.0
4.9
1.8
3.0
6.4
3.7
1.5
-4.4
3.5
5.9

4.0
1.6
5.8
7.0
4.0
2.0
.6
2.2
10.2
5.6
8.0
13.3
ENERGY BALANCE



oF/% C02
260
240
240
290
290
310
320
320
260
360
350
230
260
240
260
220
260
250
260
230
280
250
240
250
260
260
270
250
[mean]
% Deviation
From Mean
[mean] - [mean]
0.9
-6.6
-9.2
11.6
10.7
20.6
21.3
23.5
-2.1
- 1.9
-2.8
-13.2
- 1.3
-8.4
-0.7
-14.3
-1.0
-2.3
1.1
- 10.4
8.0
-5.0
-7.5
-4.3
-2.0
- 1.0
1.9
-5.7

Heat
Release
In Btu/hr
8.9
86
88
92
73
87
77
90
98
81
92
104
99
76
73
98
80
75
101
101
97
88
84
93
88
81
79
80
                                                                                                  107

-------
                                                                                                                     Table A-5.  COMPILATION'OF TEST DATA


RUN MOISTURE
1
2
2a
3
4a
4b
5
6
7.
7b
8
9«
9b
10s
lOb
lOc
11
12
13*
13b
14
WET
WET
WET
DRY
WET
WET
DRY
DRY
DRY
WET
WET
WET
WET
DRY
DRY
DRY
DRY
DRY
DRY
DRY
DRY
GRATE
1
40
35
35
35
35
35
55
55
55
55
35
50
50
55
55
55
35
55
45
35
35
SPEED ,
2
30
40
40
40
35
35
55
55
55
55
30
50
50
50
50
50
35
55
45
50
50
FPH
3
30
45
45
45
35
35
55
55
55
55
30
50
50
50
50
50
35
55
45
55
55
BED 1
1300
3780
3780
3780
980
980
915
1440
785
785
965
3160
4170
2350
2090
1960
915
785
1440
330
655
2_
2640
7170
7170
717)
4550
4550
3220
5250
4110
4110
4000
6120
7900
5100
5100
4080
4535
4535
6425
1480
7445
3_
1410
3830
3830
3830
3230
3230
3380
4500
3100
3100
3200
3260
4220
5630
10,120
1970
3375
3490
6130
1410
7030
TOTAL
5350
14,780
14,780
14,780
8760
8760
7515
11,190
7995
7995
8165
12,540
16,290
13,080
17,310
8010
8825
8810
13,995
3220
15,130
T. TOTAL
1 2 "*
25 49
26 48
26 48
26 48
11 52
11 52
12 45
13 47
10 51
10 51
12 49
25 49
26 48
18 39
12 30
25 50
10 51
9 51
10 46
10 46
4 49
j
26
26
26
26
37
37
43
40
39
39
39
26
26
43
58
25
39
40
44
44
47
OVERFIRF.
ATR. TFM

^^j^..^^

1. STOICH10METRIC CO,
1 7 1 TIVTAT £. 	
8
25
25
25
7
(,

3C
• J
3.5
JC
20
10
9


7e
. 3
2
f.

16
48
48
48
30
30
16
24
18
18
27
30
39
23
23
18
30
20
34.
10
50

9 33
26 99
26 99
26 99
22 59
22 59
15
20
14
14
22
16
21
25
45
9
23
10
5 33
10
48

35
50
35.5
35.5
55
61
80
58
77
36
59
39
75
22
102

3650
6020
6020
6020
3140
3140
3635
6580
2735
2735
4335
4850
4450
11,470
9440
4200
4215
4365
5850
2000
5850

•x.
11.0
.80
.20
.25
.30
.40
.83
. 36
.45
.49
. 24
o
.14
.48
.21
o
9.52
12.0
7.8
10.7
4.9


°2
	 X 	
8.00
20.0
18.0
19.5
19.0
19.2
18.9
19.4
19.4
19.7
21.0
20.0
20.0
19.0
20.0
20.0
11.57
7.6
11.5
8.1
12.4

ZONE

H20
	 ^
s*
15.0
3.0
3.0
3.0
3.7
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
10.6
16.4
9.9
9.9
9.4

!*?' ZONE II

CH4
1210
10
8
3.6
8.3
32
28
17
10
12
7
8
10
12
1.7
15
20
740
195
42
41

CO
PPM 	
1880
65
20
10.8
18.3
59
124
61
24
73
8
21
15
30
20
13
38
2975
3000
236
325

H,
z
— >
1650
0
0
0
0
30
88
29
10
47
38
16
19
5
42
21
56
1555
1800
14
149

co'lgl
•^3T
'Hi*
1
2.16'
3.00;,
6.13
slstf
2.fjj
(p'!
c|w
('*@
(.2T)
(0)
3.66
7. 18
8.50
5.17
5.1

02
— % 	
10.5
20.5
18.0
19.7
19.0
19.8
17.8
16.9
14.0
14.6
15.8
(20.0)
(20.0)
(19.0)
(20.0)
(20.0)
11.6
12.6
10.7
14.4
15.5

H20
	 ^.
22
3
3
3
3.7
3
5
5
9-
9
5
(5)
(5)
(5)
(5)
(5)
6.2
10.9
9.9
9.9
6.2

CH4

1333
12
8.3
4
8
32
42
53
58
68
32
(8)
(10)
(12)
(1.7)
(15)
24
201
142
16
16

CO
.PPM 	
1738
78
23
12
12
45
253
355
495
628
344
(21)
(15)
(30)
(20)
(13)
12
2750
2500
210
270

H2
-^
14
0
0
0
30
41
217
215
78
96
25
(16)
(19)
(5)
(42)
(21)
38
1250
1800
14
39

co2
4^-
7.00
2.90
1.75
3.85
1.8
2.73
3.45
2.76
4.10
2.40
.94
1.10
.22
2.40
.21
0
1.65
4.60
4.50
3.0
2.28

°2
	 •/, 	
13.0
18.5
16.8
16.3
18.0
17.4
16.4
17.1
15.6
18.1
15.9
19.0
20.0
18.0
20.0
20.0
19.4
15.3
15.9
17.0
18.4

ZONE III
V
— »
8
5
5
13
3.7
5
6
6
9
9
3
3
3
4
4
4
2.5
6.2
5.1
5.1
4.5

CH4

34.2
42.0
36.4
50.5
38.9
33.7
49.2
53.6
38.2
32.7
28.0
7. 4
3.8
42.7
45.1
51.5
31.4
39.7 .
38.2
25.0
22.0

TOTAL CFH ,
X 10°
5.27
9.67
9.27
11.02
9.21
7.95
7.58
6.85
9.04
7.73
7.74
2.71
2.12
6.58
6.08
6.79
7.57
6.18
8.07
3.86
3.32

NOTE
    :  Number» in parenthesis were estimated from general profiles.
                                                                                                                                              i;D8

-------
                                    APPENDIX B

       CALCULATION OF UNMIXEDNESS FACTOR OF REFUSE BED OFFGAS

     The ratios of moles of various components with respect to moles of offgas are shown in
 Table B-l below:
                                Table B-1. MOLE RATIOS
              Unreacted Mixture                           Reacted Mixture
              	Moles    Percent                  	Moles     Percent	
                                100%               .                   100%
""--                  *                  -Combustion*
^^ * •;!*•••                  •                   ^»wi i ihsut*viwi i            i    	           	
                          1 + 4.76nQ2 + nx   Products                 1 + 3.76nQ2 + nx

                          (4.76nQ , • 100%                          3.76nQ  • 100%
Stoichiometric        4-7fin0	 "2               3.76n	
 Air                      21+4.76n02+nXs                        2  1 + 3.76nO2 +
    U2   As                                 02
n
                             nx  • 100%                               nx • 100%
Excess Air             nx	— Excess Air          n	
                       S   1 + 4"76no2 + "xs                     X$    1 + 3'76no2 + \

Total         1+4.76nn  + n       100%      Total       1 + 3.76n  +n       100%
                     °2   xs                                O2   Xj
 Note that in a Stoichiometric mixture nx  is zero. For one mole of offgas, n02 moles of
 oxygen are required and therefore 4.76 • n0 2 moles of air. The mole fraction of offgas in a
 Stoichiometric ratio is therefore 4.76 • n0 2:
                          moles of combustible          1
                     Ys~      total moles      ~  l+4.76-n02

 The mole fraction of offgas in a typical mixture (not Stoichiometric) is:

                               y =	1	   .                        (B-2)
                                  l-r4.76(n02+nXs)

 To relate these relations to the gas analysis, consider a blending of a Stoichiometric mixture
 ys • Ct; the concentration of offgas in the excess air is zero. Since concentration, C, can be
 calculated from the mole fraction and the ideal gas law, then:

                                 C=y   -C=y^
                                 ^s   y&  M  y$
                                 C=y -C,-,
                                                 PT
                                                  p                            (B-3)
                                                 —
  'None of the combustion reactions involved results in a net production of mols.

                                         109

-------
For blending,


  Cs (volume of stoichiometric mixture) + O.D (volume of excess air) = C (volume of total)


Hence:


                      (y, ' Ct)V, = (y • Ct)Vt or ysns = ynt.                    (B-4)


Then, for the example in Table B-l,

                             PV
                             PV
                           =-- = n = l+
                                                                            (B-5)
Therefore:

                          y   =  ns   _    1 +4.76 'n02

                         7T." "»T  = l+4.76(n02+nXs)                    (B'6)


     The initial concentration or mole  fraction of offgas is not always known, and it is

important to determine y/ys in terms of gas analysis after combustion. The same approach is
used as before, considering a mole of final combustion products rather than offgas. In this
case:
and
                                      nco2
                             %C02|S = -  -100%                          (B-7)
                                     nco2
                             %CO2 = - -  -100%
                                       nt
Hence

                        <%ro.     nc«           1 + 3.76 • n.
                                                                           (B-8)
%C02    %          l+3.76-n92
                       %CO2|S    nt      1+3.76-n02+4.76 nXj
                                        10

-------
This can be rearranged to:



                 %CO2 = %CO2 ls
                                      (1+3.76 -n02)
                                  U+4.76(n02+nXs)-n02
                 %C021
                          (l+3.76-n02)y
Solving for y yields:
Then:
(%CO2/%C(
_l+n02(%C02,
%C02
%C02|S

M) ir i
'%C02|J |_l+3.76n02 _
r / i %C02 \n
\l+3.76n0, %C02|S/
1 ' " %co2
%CO,L n°2
When n0  is small, the equation reduces to
      °2
                              Ay =
                                         _%C02_\
                                         %co2ij
Example: For the bed offgas where CO/CO2 = 0.5, n02 = 0.058, ys = 0.784


    At a temperature of 1500°F, %CO2 = 7.2%.

                                      '_J	} -  2,
                            ]+ao58.  0*3.76(OW    IS-
     Ay = 0.783
                1 -
                     I./.70
                    15.1%
                                                           .1%
                                            '*
                                  Ay = 0.404




Using the simplified expression above, Ay is estimated to be 0.410.


    The required oxygen, O2 req, is calculated based upon CO:


                            02req = n02 • (CO/CO I s)
                                                                         (B-9)
                                                                        (B-10)
                                                                        (B-ll)
                                                                        (B-12)
                                                                       (B-13)
                                                                        (B-14)
                                       11

-------
    The available oxygen is equal to %O2  from the gas analyses.

Hence:
                             02 req       n02 (CO/CO|g)
                           O2 available        %02

    For the example above, a CO concentration of 100 ppm results in a value of 8.6 x 10~4.

    The factor >/c'  /Ay is obtained from Figure 3 and is approximately 0.40.

    Since

                             ^fif = (jfF /Ay) O • Ay                      (B-l 6)

    then, the unmixedness factor for the example is:

                               0.40x0.404 = 0.162
                                       112

-------
                                  APPENDIX C

             MATHEMATICAL CHARACTERIZATION OF QUENCHING

    The problem of calculating the rate at which a pocket of fuel will be quenched by
reduced temperature and determining the corresponding fuel concentration at the time of
quenching may be modeled as described below:

    The temperature  of  the gas is cooled by conduction of heat from the fuel at the
adiabatic flame temperature to the  surrounding air at ambient conditions. This represents
the most rapid rate of conduction  heating and  other factors, such as air  preheating or
radiative heating, will diminish this rate as well.
    Consider a portion of fuel as shown in
the figure right  in which the temperature is  AIR
initially Tf. At  time,  t = 0  the  surface is
brought to temperature T0 and held constant;
again  this  is the  maximum  gradient.  The
governing boundary value problem is:
     dT/at  =  a a2 T/a/2

         T(0,z) =  Tf

         T(t, 0) =TO,
        T(t, oo)  =  T0
The solution is:
              Tf-T0
                      =  l  ~e rf
     If this is evaluated for gases at the quench temperature, then:
           Tf   = 3100°F

           TO   = 100°F,and
           Tq   = 1100° F (quench temperature).
     and, therefore:
         erf
=  1 -(3100-1100/3100-100)  Si 1/3
                                                 (C-l)
                                                 (C-2)
                                                                           (C-3)
                                        13

-------
where zq(t) is the distance from the initial surface to where the reaction is quenched. The
rate of propagation of the quenching front is:
                               d(zq)/dt =  0.305 v/^/T                        (C-4)

     If  one assumes that the fuel  is initially at the adiabatic flame temperature (less a
reasonable  amount of heat loss) and must be cooled by dilution with ambient air, then the
change  in  temperature and  concentration will be continuous. For any combustibles to
remain following this transition, the mechanism for cooling would have to precede so that
the stoichiometric concentration Cs would  not be attained before the  fuel reached the
quench  temperature. The above analysis shows the rate of propagation of the quench front;
the propagation of  the  plane where  Cs  is reached must be slower than 0.305 \/a7t
Otherwise,  a condition would exist in which  fuel and air would exist in proper proportions
and above  the quench temperature; the expected result of  this  condition is a  complete
burnout of fuel.

     First,  consider the case  of air diffusing without reaction, the analogous case to the heat
condition problem. The governing equation for Ca is:

         aca/at = Dao2ca/az2)         ca +cf  =  cx

         Ca (t,  0)= Cao,Ca(t,oo)  = 0                                       (C-5)

         Ca (0, z)  =  0

The solution is:

              (Ca0 - Ca/Ca0)=  erf (z/ V^B                               (C-6)


For the  case where CO/CO2 = 0.5, Cs was shown to be 0.78.

Then:
         C.  -C.
                    =  0.78  = erf (^4 Da t  >                                (C-7)
                                       a
           C
           U
in which case
                         zs =  0.87 >/4D,t                                  (C-8)
                                        114

-------
  Then, when
             a = 11.2 ft2 /hr (constant)
                           D=  9.3 ft2/hr (constant)

                             d(zs)
                              dt
                                    = 0.79V/07T
      AIR
                   FUEL
      -z
                   z=o
                                          +z
                          fuel + n  'air

  The governing equation for each region is:

                     Air
               =  C
o2 c
                          -°o Tq is a certainty and
       no combustible would be expected.
  \
            Consider  next  the  effect of  a
       rapid irreversible reaction as shown in
       the figure left. Since both air and fuel
       cannot co-exist, there is  a plane at z
       where  the  concentration  of both is
       zero. For the constant molar case:

   (1 + n) products
                                                Fuel
                                                cf + CP = CT
3 Cf/d t = Df (32 Cf/9z2) z' < 7 < oo

Cf(0,z')  =CfQ

Cf (t, z') = 0, Cf (t, oo)  = cffl       (G9)
                                 z = z
  The solution is of the form
                         = aj + a2 erf
                      Cf =bj + b2 erf[-
                                                                              (C-10)
                                           115

-------
Solving for the boundary values, the solution is:
C.  -C
                                    l+erfU/4Dft
where
                                    1 -
                                    z =
                                                                            (c-ii)
From the final boundary  conditions, representing the reaction stoichiometry, the implicit
relation for TJ is derived:
             T? Df)    1? • Cf

For the case Df = Da = D, Ca  = Cf  =1 (pure fluids) the relation reduces to
          erf
                                                                            (G13)
Then for CO/CO2 = 0.5, n = 0.23 mole air/mole fuel
                               dt
             . (Z') = 0.57 V57T
                                                                            (G14)
Because of the small amount of air required, the value for T? is still not small enough for the
rate of propagation of Cs to be slower than that for T .
                                        116

-------
     The final step in the analysis is to consider the effects of a fuel temperature lower than
the theoretical adiabatic flame temperature. Such an occurrence results from two phenom-
ena. First, the combustion reactions in the bed do  not go to completion so that all of the
heat of reaction is not liberated and the flame temperature is proportionately lower. The
additional amount of heat is liberated as oxygen diffuses, but the net reduction in the heat
liberated would  occur if a pocket of combustibles  were quenched before the reaction was
completed.

     A second  mechanism  for having a fuel  temperature lower than the  adiabatic  flame
temperature is a result of heat transfer from  the  hot combustion gases to the refuse bed
itself. This is evidenced by an increase in the temperature  of the refuse bed as well as an
increase in the production  of gasification or  pyrolysis products.  In  the case of pyrolysis
which  produces large amounts  of hydrocarbons, the consistency  with  the gasification
premise for bed-burning is largely the result of the decomposition of methane to carbon and
hydrogen, which is substantially complete at temperatures  above a 1000°F.

     Finally, a pocket of fuel can lose heat through radiation. This mechanism tends to cool
the hottest gases and heat the colder gases and will result in a net loss of heat for the high
temperature fuel pockets.

     If we assume that the  temperature of the fuel is only 2000°F, as a result of the three
mechanisms above,  then  the left  hand side of Eq.  (C-3) has the value of 0.48 rather than
0.305.  Hence, the speed of propagation of the quenched  plane would be greatly increased;
but  in this particular case, the  increase would still not be sufficient to overcome  the
propagation of the Cs plane. However, the assumption of pyrolysis on a larger scale suggests
that the ratio CO/CO2 could be much larger thanO.5. Hence, the factor n in Eq. (C-l 3) would
be larger and the rate of propagation for the C plane would be  correspondingly slower. As
an  example, consider the ratio  CO/CO2 =1.0. From a  material  balance  the mols  of air
required practically  double  so that n = 0.5. Then the left hand side of Eq. (C-13) has a value
of 1/3. The solution for this value is identical to the initial  example for the proportion of T
Eq.  (C-13). Under these conditions the propagation of Cs would occur at a  slower rate than
the propagation of the quench plane. When this condition is obtained, Eq. (C-9) would still be
an adequate description of the process if the boundary conditions were changed to:

                                   Cf(t,z.)-C,
-D_

                       = n
-D,
                                                      ac
                                                           Z=z  -
                                                              q  -J
                              (C-l 5)
                                         117

-------
The concentration profiles are shown schematically below:
where in this case z  marks the location of the quench plane. To the right side of z the mix-
ture is above the quench temperature and, therefore, the oxygen concentration must be 0. To
the left side, the mixture is below the quench temperature so that the co-existence of both
oxygen and fuel is possible. Under these conditions the equation for Ca is unchanged but
the equation for Cf becomes:
     Cf  -C,
       fo     f
     C  -Q,
                     1 -erf
V/4D7T
                     1 -erf
              Z-  < Z < +
                 erf(zV4Dft)
                                          — oo < z < Z,.
                                                (C-16)
Interestingly  enough, the concentration  where quenching  occurs is constant  as  can be
deduced  from Eq. (C-10)forCf by substituting zq  = .305 V4a t. If for the mixture on the
right hand side of z  the diffusion equation of the form for Eq. (C-5) is solved for fuel and
products individually as well as the two combined, then it can be shown that
                                      =(cp/cf)|z=
                                                (C-17)
The stoichiometric boundary condition in the case of quenching reflects the net consump-
tion of fuel at the boundary z  . This can be solved to yield:
     1 -erf
     1 4-erf
                Ja.
                          n(C   -C
                                ao
                                                    exp
                               4Dat
                                            4Df t
                                                        1 -
                                                (C-18)
                                        118

-------
which is similar to (Eq. 12), .but with the substitution of (Cf  - Cn) for Cf  and with the
                                                        lo     i      'o
additional  term [1 - (Cq/Cf  -Cq)]  representing  the  flux of fuel into  the  quenched
mixture. Equation (C-l 2) was used to solve for z  . In this case, however, z  is determined
from the heat conduction problem. Substituting Eq. (C-3), for example, and assuming equal
diffusivities (Da = Df = D), Eq. (C-l 8) can be used to calculate the quench concentration C .
Simplifying the above,
*(•
                                               erf (0.305 ^ ,
                                             1 + erf (0.305 V^/D ')
     C  = — 1C
     S     2  lfo
which reduces to Eq. (C-l3) when C   equals zero and Cf =Ca  = 1.0. For the case when
n = 0.5, C   has a value of only 0.034. If, however, z is assumed to equal 0.48^40; t, as
suggested above, then for n = 0.5, C is equal to 0.21 which is substantial.

     Up  to  now analysis has centered around showing not necessarily that quenching will
always occur, but only that  under  the  right set of  circumstances it can occur. The
experimental data presented  in  Section VII clearly show that in all three zones a certain
amount of quenching  is witnessed. Several additional sophistications can be added to the
analysis,  including the  effect  of differences in diffusivities of fuel and air and the effect of
considering  the  diffusivities  as functions of  temperature  rather  than  as mean average
constants. The latter is of particular  note because diffusivities vary with T3/2, while the
conductivity is a function of T1/2. The basic conduction and diffusion equations reduce to:
            Conduction                                 Diffusion

           d6       _  ..    d20           '      dC   _   ^     d
     d     di?               drj2                 dTj            drj
Note that both the  thermal diffusivity a and the molecular diffusivity D are functions of
T3'2, but because of the placement of the derivatives the differential equations are not of
the same temperature functionality and the respective solutions will reflect this difference.
Unfortunately, these equations are all highly non-linear and require a numerical solution
rather than  an analytical  one. Since little could be gained from this additional degree of
sophistication, the numerical analysis was not carried out.
                                         119

-------
     From the above analysis, it is clear that  the quenching phenomena can cause com-
bustible emissions under the right set of circumstances so that a final refinement is required
to devise a method for calculating these emissions. In reality, the transfer between the fuel
phase and  the air phase cannot be legitimately represented by semi-infinite medium since
both of these are finite. The solutions of the basic differential equations with respect to
finite boundary conditions have been worked out and will not be attempted here.

     The simultaneous heat and mass transfer mechanism is not the only means by which
emissions  can  be  produced in gases  having the appropriate  amount of air  and sufficient
temperature to sustain combustion. In a recent paper by Sivashinsky and Gutfinger* the
hypothesis that emissions can be caused because of the finite speed of chemical kinetics was
explored. They concluded that during the course of combustion not all of the combustible
gases are consumed in the reaction zone and that a portion of the combustible component
passes through the flame front without reacting with oxygen.

     Still others have suggested a stochastic model based upon eddy collisions and applying
the same type  of probability equations used in the kinetic theory of gases.

    The exact mechanism  or combination of mechanisms is not known and research is
underway  in this area. The quantification of quenching cannot proceed beyond the level
given here until a better physical description is at hand.
    Sivashinsky, G. I.,  and Gutfinger, C.,  The Extinction of Spherical Diffusion Flames, Paper B5,
    Proc. 1973 International  Seminar  on  Heat  Transfer  from  Flames, Aug., 27-31, 1973, Trogir,
    Yugoslavia.
                                         120

-------
                                   APPENDIX D

BASIC PRINCIPLES OF COMBUSTION

     Within  this appendix,  the  fundamental principles of  combustion are presented for
those unfamiliar with combustion chemistry.

Reaction Chemistry

     The principles of combustion of refuse are no different than those of other fuels, and a
great  deal of  study  has been  centered  around the reactions that occur (3, 4). Most
hydrocarbon fuels are parafinic  and can be represented by the formula (CH2)n. Refuse is
more cellulosic than parafinic and a more accurate formula would be CHn (H2 0)m . The
equation for the complete combustion of refuse can  then be given as:
                                                                H20.         (D-l)

     In reality the reaction does not go to completion for some fraction of their fuel and
off-gas products. Partial combustion products, such as CO, H2, hydrocarbons, and soot,
often result. The step reactions which take place either in the refuse bed or overfire region
include oxidation reactions:

              C   +1/202    < - >   CO                                    (D-2)
              C0+l/202    < - >   C02                                   (D-3)
              H2+l/202    < - »   H20                                   (EM)

     and reduction reactions:

              C  +C02       < - »   2CO                                   (D-5)
              C  +H20     . < - >•   CO + H2                               (D-6)

     By combining reactions [5] and [6] , one obtains:

               CO + H2O      < - '—+   CO2+H2

Tlie  last  of these reactions, known as the water gas shift reaction, is  of considerable
importance in describing the gaseous equilibrium in bed-burning processes.
                                        121

-------
Reaction Kinetics

     The reaction  rate for a particular molecular species is dependent upon a combination
of all the above reactions. For example, the rate of reaction for carbon monoxide is given by
the equation:

                                       (r4-rs+2r7+r8).                       (D-8)
                                                         .
                                dt
A similar equation  could be derived for each  of the  constituents, provided that all the
occurring reactions involving the constituent were identified.

     The expression for  the  chemical reaction  rate is also dependent  upon the specific
mechanism of reaction. In very simplified terms, the rate can be expressed by the law of
mass action applied to molecular reactants. For reaction  [D-3]  above:

                                 r3=k3(CO)(02)'/z                             (D-9)

where k3 = chemical kinetic constant, and (CO), (O2) = concentration of molecular species.

     Unfortunately,  the mechanisms for  combustion reactions are rarely as simple  as the
elementary expressions above. Typically, the mechanisms involve such things as free-radical
chain reactions or other types of activated complexes. The exact mechanisms have been
worked out over years of research and will not be  discussed in detail here.

     All of the oxygen-consuming reactions have extremely rapid reaction rates. This has
been  experimentally verified in many instances where reactions were shown to  go to
completion within a matter of microseconds, and certainly the rate of reaction in an internal
combustion engine,  or in  a hydrocarbon explosion, attests to the rapidity of the chemical
kinetics.

     The above reactions can also  go in the reverse  direction, and  although the  same
principles apply in determining the chemical kinetics  for these reactions, the rate constants
are different. The ratio of the rate  of the forward reaction  to  the rate of the backward
reaction determines  the equilibrium of the reaction  which is discussed in  detail in a later
section.

Activation Effects

     All the kinetic  expressions have their bases in statistical  mechanics and represent the
rate at which molecules or other reacting species will collide in such a way as to undergo a
chemical change. But molecules must not only collide  before a reaction takes place, they
must also have sufficient  energy to  be able to  rearrange chemical bonds  to undergo the
                                         122

-------
chemical change. The large majority of molecules colliding do not have this energy, and thus
the collision will  not result in a chemical reaction. The process is shown schematically
below:
                   Energy
                              Reactants
Products
The  average molecule  is assumed  to have  an energy level represented  by point A. Upon
collision, the molecule does not react.  However, certain molecules, because of past history,
will attain energy states higher than average. When one of these reaches the energy  level of
point B, then, like a roller coaster, it will seek a point of lower energy and will either return
to the normal state A, or react upon collision to form products having an energy represented
by point C.  The height of the energy barrier Eais  the activation energy, and the difference
between points A and C is the net energy of reaction, the latter being directly related to the
heat of combustion.

     For both  the  forward and reverse reactions,  the  dependence  of the reaction rate
constant was shown by Arrhenius to be exponentially dependent upon the  activation energy
as shown by the equation:
                                                                               (D-10)

where  Ea for the forward reaction is the difference between points A and B, while Ea for
the reverse reaction is the difference between points C and B. In the case of the exothermic
reaction, such as combustion, heat is liberated and the forward-reaction activation energy is
less than the value for the reverse reaction, so that the rate of forward reaction is faster and
the net release of energy allows the reaction to be self-sustaining once it is initiated. This
suggests an additional ramification of the activation energy — the concept of ignition.

     Consider a combustible mixture in which none of the molecules has sufficient energy
to exceed the energy barrier for a given reaction. In this case, none of the molecules will
undergo  a reactive collision and combustion will not occur.  As the temperature of the
system  is increased,  the molecules  increase in energy to  the point where one of the
molecules will  gain sufficient energy to exceed the energy barrier.  When this happens, a
combustion reaction will occur, liberating energy, thereby increasing the temperature of the
gas, and causing additional molecules to exceed the energy barrier. The chain process repeats
again and again, and the temperature of the  gas increases dramatically. The temperature at
which this series initially begins is known as the ignition temperature.
                                         123

-------
Chemical Equilibrium

     Because the chemical reactions are occurring simultaneously in both the forward and
reverse directions, the reaction will ultimately never reach the state of "complete combus-
tion." The equilibrium between the forward and  the reverse reactions is determined from
the ratio of the respective rate constants.

     In  the case of the oxidation reaction, the heat of reaction is very high and the forward
reaction occurs at a much faster rate than does the reverse reaction. The reduction reaction,
on the other hand, has a low heat of reaction and is, in some cases, endothermic (requires
heat). As a result of the Arrhenius Equation [10] , the equilibrium constant is a function of
temperature according to the relation:
                                                                             (D-ll)
This relation is shown for the various combustion reactions in Figure D-l .  In the case of the
oxidation reactions, reaction [3] , for example, the equilibrium between carbon monoxide,
oxygen,  and carbon dioxide at 1800°F is  on the order of 107 atm"1  . For a chemical
equilibrium-controlled reaction  in which the stack gas contains  12% CO2 (dry) a material
balance suggest approximately  8% oxygen. The  concentration of carbon monoxide  is
calculated, below:
                      K4 = 107atm-'/2 =    (%C°2)
                                      (% CO) (% 02 )1/2   ( 1 atm)'/2           (D- 1 2)
                                      (% CO) = 0.04 ppm
     For reaction [7] above, the equilibrium at 2000° F is equal to:
                                                =20
                                 7   (CO)(H20)
Diffusion Flames
     Unfortunately, the overall rate of chemical reaction cannot always be controlled by the
chemical kinetics and equilibrium is often unattained. The controlling step often is the rate
at which fuel and oxygen can be brought together to react.  As a good example, the rate at
which soot particles burn is controlled primarily by the rate at which oxygen diffuses to the
surface of the soot, rather than by the rate at which carbon (soot) and oxygen combine to
form carbon monoxide. Flames of this type are diffusion flames and are distinguished  from
premixed flames in that oxygen must diffuse into the fuel to form a combustible mixture.
Examples of  diffusion flames  range  from  candles  to bonfires.  A  distinction is made,
however, in the type of diffusion flame, depending upon the type of diffusion occurring.
For the small candle, the oxygen is fed in by molecular diffusion and the  flame is called a
laminar diffusion flame. In the case  of  a bonfire, the  hot combustion gases have enough
buoyancy to generate a considerable amount of turbulence and draw air into the base of the
flame by a convective transport mechanism. These flames are turbulent diffusion flames.

                                        124

-------
   o


   8.8
   O O
   ^-« r^.
                                 r,°F
         in  t m
o   o    o o  o  o    o
    §o    o o  o  o    o
    in    o oo  u>  i-    CM
ro   eg    ox —i  —<  —i    -H
-5
 0.0002   0.0004   0.0006   0.0008   0.0010   0.0012   0.0014  0.0016



                                T' K
      Figure D-1   Equilibrium constants of combustion reactions.
                                   125

-------
     In a laminar diffusion  flame the combustion is controlled by molecular diffusion
described in generalized coordinates by Pick's Law:
The rate of flux  toward the  flame boundary is  directly related to the difference  in
concentrations  at the boundary and the concentration at a distance from the boundary, i.e.,
the  concentration  gradient.  The concentration profile  about the flame front  can  be
described by a mathematically continuous function analytically determined by the simultan-
eous  solution  of Pick's Law and the law of conservation of mass for  a given system
geometry. The  solution is relatively straightforward and has been discussed at great length in
the scientific journals.

     In a turbulent system the solution for diffusive transport is not so easy a problem and
the mathematical theory is still evolving around this issue. Unlike the laminar problems,
turbulent transport  takes place in bulk rather than on a molecular scale, i.e., groups  of
molecules tend to transfer together  from  one place to another via a convective transport
mechanism. The process is referred  to as  eddy  diffusion. In a turbulent regime, a whole
range of eddy sizes exist and there are eddies within eddies. An adequate description of the
eddy  characteristics requires statistical techniques of a much greater sophistication than the
analytical techniques applied to the laminar diffusion process.

     The incinerator does have  its own air supplies  and can best be described  as partially
premixed turbulent diffusion  flame. However, the premixing notwithstanding,  the rate  of
gas phase combustion is controlled  by turbulent diffusion  and the principles describing
turbulent diffusion flames will apply.

Material  Balances

     In several  studies  of the nature  and composition of refuse, including those  specifically
related to the City of Newton's  refuse, the average composition of dry ash-free refuse was
approximately CH0.4 (H2O)0.6 so that the  combustion reaction is:

                   CH0.4 (H2O)0.6  + 1.1 (02)   — >CO2+0.8H2O               (D-15)

The coefficients of each reactant and product indicate the relative amounts (in mols) that
are required to combine  exactly during the reaction. These proportions can be converted  to
actual compositions or weight of burning material using the law of conservation of mass. A
summary of the relative proportions of combustion products is shown in Table D-l. More
often the composition  of a gas sample is given in volume percent which is equivalent to the
mol percent shown in the table as well. Another unit, the Orsat analysis, is usually given on
a dry  basis. The interrelationship between these various quantities can be summarized by the
equations below  applied for carbon dioxide balances. These equations, Ns represent the
                                         126

-------
number of moles of gases in a stoichiometric mixture containing CO2, H2O,  and the
nitrogen equivalent of oxygen required to complete combustion; i.e.:
                           +nH20+nN2,s
                                                (D-16)
If nxs  represents  the  number of moles of oxygen present in excess of that required to
complete combustion, then the total number of moles, nt, is determined from the equation:

                  n,  = n,. + 4.76nx<,
                   la       AS

    The gas composition (is mole percent) can be calculated for CO2 as follows:

                                  n/-./-v
                        C02)W =
                     (%C02)d   =
                                   n
                                     'CO-
                          X 100%
                              *

                              X  100%
                                                (D-18)

                                                (D-19)
                                   IV n
                                 _  CO 2
                            's,w
                                          X  100%
                                                (D-20)
                        >CO,)
 2's,d
           nco-
                                   ns~nH20
                               X 100%
(D-21)
     Since  1.1 moles  of oxygen are required per mole of carbon (Equation [D-l 5]), the
percentage of excess air is calculated by the following equation:

                                              (%02)d
                (%XS) =
n
                           xs
                        1-lnco-
       X 100% = -
                  1.1 (%C02)d
                                        X 100%
(D-22)
 Likewise, if one assumes that the refuse initially has 0.7 inch   H2O per mole of C (35%
 moisture on an ash-free basis):
ns=nco2 +
          +3.76 (l.
                                                    = 6.64 n
 (D-23)
     Finally, Equation [D-17] can be rearranged to the form:
                                  ns    4.76 n
                                             xs  _
                                                 = 1
                                                 (D-24)
     Combining Eqs. (D-23) and D-24) and rearranging the fraction so that nt and ns  are
both denominators yields:
                                       127

-------
                               n,
                                CO-
                               nco-
                              6-64nco2
  4.76
                                                       = 1
                              (D-25)
or
                                  15.1%
(%02)d    =
 21.1%
                                                                            (D-26)
                   Table D-1.  PRODUCTS OF REFUSE COMBUSTION
                  Fuel
  Air
Refuse
C
H
0
Moisture (H20)
Total
Component
C02
H20
N2
Mols
1.0
1.6
0.6
0.7

Mols
1.00
1.50
4.14
6.64
02 N2
(required) (3.76 XO2)
1.0 3.76
0.4 1 .50
(0.3) (1.12)

1.1 4.14
Combustion Products
%Wet % Dry Weight (%)
15.1 19.5 23.5
22.6 — 14.5
62.3 80.5 62.0
100.0 100.0 100.0
Energy Balances

     In  an adiabatic  process (no external heat loss or gain to the system), a stoichiometric
mixture of refuse and air reacts liberating approximately 7500 Btu/lb of dry ash-free refuse.
All of the liberated heat goes to raising the temperature of the combustion products from
T0 to the adiabatic flame temperature Ta. The energy equation is:
(nsCp )(Ta-T0)=  Z  (nCp)i(Ta
                                                •T0) = 7500 Btu/lb
                               (D-27)
                                        128

-------
where:
     HJ   = mass of component i per Ib of dry ash-free refuse
ns   =
           n = l
               n; (stoichiometric mixture)
     Cp .  =  heat capacity of component i
The heat capacity for the stoichiometric mixture is shown in Table D-2.

                             Table D-2. HEAT CAPACITIES
   Component

       C02
       H20
       N2

        Total
             Mols

             15.1
             22.6
             62.3

            100.0
(cals/gmols —'

     13.0
      9.0
      7.4

      8.61
                                                                     196.3
                                                                     203.4
                                                                     461.0

                                                                     860.7
     The final temperature of the stoichiometric mixture  is referred to as the theoretical
adiabatic flame temperature and is approximately 2950°F for refuse. A temperature of this
intensity can  damage  a refractory  incinerator. The primary method for reducing the
temperature to a reasonable operating range is through the addition of excess air. The gas
temperature is lowered because the heat liberated by  combustion is  used to raise the
temperature of more than just the combustion products,. The energy balance is as follows:
         (n, Cps) (T -T0) + (4.76 nxs C?xs) (T -T0) = 7500 Btu/lb
     Equating Equations D-27 and D-28 yields:

                                 "sCPs
T-T
*a   1o
                     n.C  •+4.76nv«C
                                                            'xs
     Equation [D-29] can be simplified to:
                      T-T«\      4.76 n
                     T-T
                     A   l
                                         xs
                                 ns
                                 ~Pxs

                                 :Ps
                                                       AH/(T-T0)
                T-T
                1   io
               L T, ~ Tn
                                                                             (D-28)
                                                                             (D-29)
                                                                             (D-30)
                                         129

-------
     Substituting Equation [D-26]  and rating that nxs/ns equals (%O2)/100%> the equation
becomes:
              T-T0\     (%CQ2)W   ,   (%02)        p.jpx.      T-T0\
              Ta-T0)    (%C02)S>W    (21.1%)          Cp        \Ta-Tj
The second term on the right is a correction factor for the changes in the heat capacity in
going from reactants to products and is on the order of 0.1. Since both the stoichiometric
CO2  and the adiabatic flame temperature are fixed for a given refuse, the equation predicts
a linear relationship between temperature and C02  for a given refuse.
                                       130

-------
                                   APPENDIX E

              CHARACTERIZATION OF THE BED-BURNING PROCESS

     The overall objective of developing a model of the processes  occurring within  the
refuse bed is to determine the optimum distribution of both overfire and underfire. Previous
attempts at describing the bed burning have approached the burning as a combination of
several processes. The resulting models have been simplistic, providing little more than a
qualitative description of the refuse burning process.  The difficulty does not lie in under-
standing the physical processes that are occurring, but rather in mathematically combining
the processes into a single coherent model.

     The approach taken here is somewhat different. In the  mixing study several phenom-
ena  which could occur and control the bed-burning mechanism  were identified. The
experimental data on burning obtained in this program allow these phenomena to be placed
in perspective so that conclusions regarding design  relations can be drawn.

     Following a series  of  short tests over each of  the  three  grates of the Newton
incinerator, we selected the first three sample ports for complete bed tests. The data from
these tests are given in Table E-l. These two tests  and the tests made in the breech form the
basis for conformation of the bed model presented here.

DESCRIPTION OF THE BURNING PROCESS

     Several processes occur within the fuel bed, including drying,  ignition, gasification,
combustion, and char burnout. Earlier qualitative models for  fuel bed combustion identified
separate drying,  gasification, and char-burning regions defined so that a single process  was
occurring in each region. However, these operations actually occur simultaneously because
of the heterogeneous character of refuse.

     The parameters controlling the burning bed  are not well defined. Certainly one of the
more important parameters is air availability  which is  dependent upon how much and from
what region air is introduced into  the furnace. Most municipal incinerators are designed to
introduce the primary combustion  air from beneath the refuse bed, i.e., underfire air.

     Consider  a fuel bed supplied  exclusively with underfire air. Oxygen will be consumed
as the underfire  air passes through the bed so that,  for underfire air rates less than that
required for complete  combustion,  there will occur a point within the bed where  the
concentration  of oxygen becomes zero. In a recent  controlled study of incineration the
existence of this point was verified.*
                                        131

-------
                          Table E-1.   TEST RESULTS OVER BED1

                                       Sample Port
      Bed Test 1
  1
                              Average
      C02
      02
      CO
      H2
      HC
      H2O2
      H20/C02
      C0/C02
      Kwg
      Temperature

      Bed Test 2

      C02
      02
      CO
      H2
      HC
      H20
      H20/C02
      C0/C02
      Kwg
      Temperature
 5.93
 13.4
 3.2
 1.72
 1.16
 11.51
 1.94
 0.54
 3.61
1835
 13.3
 6.0
 3.35
 2.7
 1.94
 12.9
 0.97
 0.25
 1.20
1787
  1.31
 18.9
  0.26
  0.21
  0.01
  2.15
  1.64
  0.20
  2.03
1905
 6.4
14.2
 1.00
 0.77
 0.34
19.4
 3.03
 0.16
 3.94
1758
  5.8
 14.6
  0.13
  0.09
  0.03
  8.84
  1.52
  0.02
  2.2
1928
 12.0
 8.6
 0.26
 0.19
 0.39
 26.4
 1.70
 0.02
 2.33
1811
   4.35
  15.63
   1.20
   0.67
   0.40
   7.50
   1.72
   0.28
   3.1
1889
  10.6
  9.6
  1.54
  1.22
  0.89
  17.57
  1.66
  0.15
  2.09
1785
      1. Complete sets of data are limited due to experimental difficulties directly over the bed.
      2. Calculated from national balance.
     In our tests, this was not the case.  In both tests directly over the bed, the oxygen
concentration was on the order of 10%.

     The  presence of air  in the bed offgas could be  the  result of any one of the three
mechanisms shown in Figure E-1.
                                           132

-------
           ENTRAPMENT
                 CHANNELING
                                                                     EXCESS AIR
 OVERFIRE
    AIR
OVERFIRE
   AIR
                                      W////A
                                            REFUSE /
                                          V,
          UNDERFIRE AIR
              UNDERFIRE AIR
                                                                    UNDERFIRE AIR
                           Figure E—1   Oxygen breakthrough mechanisms
Excess Air
    When  the refuse bed burns to a point where the kinetics for the solid combustion
become very slow, then air can pass through the bed at a faster rate than that at which it can
be consumed. When the amount of excess air is small compared to the amount of air that is
consumed, the effect is of little consequence, since the air breaking through the bed will be
well mixed with the partial combustion products, and complete burnout of the complete
emissions will be eminent. However, when the fraction of oxygen breakthrough is much
larger than the amount of oxygen that can be consumed by the reaction, the heat liberated
by combustion will not be sufficient to raise the  temperature of the gas stream and
quenching will occur. The latter situation  is of particular importance in the final stages of
refuse  burnout in which the refuse  approaches a fuel-limiting condition.  In  the latter
fuel-limiting case, the  gases are uniformly mixed  and the equilibrium is  controlled  by
oxidation. In the case of both channeling and air entrainment in which non-uniform mixing
exists, the water gas shift equilibrium  will apply to those portions of the gas which contain
combustibles but  no oxygen. The non-uniformity of mixing is not  unreasonable when
considering the mechanism by which the oxygen breakthrough occurs.

Air Entrainment

    Several recent studies have shown that air can be drawn in from the overfire region and
entrained into the gases  emanating from the refuse bed. In such a case, the bed-burning
process must be considered as a combination of both an underfire and an overfire mecha-
nism in which the zero-oxygen planes by each mechanism occur within the bed.
                                       133

-------
Channeling

     Because of the lack of a completely random packing of the refuse, channels, or "blow
holes," may develop. The "blow holes" allow the air to bypass the refuse completely. This
phenomenon is a common occurrence in packed bed equipment.

     Each one of these three mechanisms could control the burning. In the case of the tests
at Newton,  the overfire entrainment mechanism was believed  to be the controlling  factor.
This is best illustrated by considering the magnitude  of local burning rates.

BURNING RATES

     The overall burning rate can  be determined from the amount of oxygen consumed,
from the  amount  ofCO2,CO, and hydrocarbons  produced, or from the amount of heat
liberated.  Table E-2 summarizes the overall burning rates derived from both carbon and
energy balances. The averages — 54 and 67  lb/hr-ft2 of grate area — based upon carbon and
energy balances, respectively,  are close to the value of 60 lb/hr-ft2 often used as a design
value. Several assumptions on bed density, ash content, moisture content, and heating value
were introduced in the  calculations. Errors in these assumed values may account for the
difference in burning rate obtained by the carbon and energy balances.

     The fraction of total air consumed which could have been introduced underfire is also
given in  Table E-2. Only one third of the total consumed air could have  originated as
underfire  air. In  this case,  the computed  burning  rates  are 1.7 to 2.1  times that corre-
sponding to stoichiometric combustion of all of the underfire air.

     The overfire entrainment mechanism is  the only one  which  would result in burning
rates higher than the burning rate equivalent to complete combustion of the underfire air.

     The results presented in  Table E-2 also indicate that  the overall burning rate is not
drastically affected by the underfire air rate, but remains relatively constant at about 60
lb/hr-ft2  of grate area in agreement with empirical design rules used for years in designing
burning surfaces.  The rule holds because  the total amount of air tends  to remain constant.
Reductions in underfire air are offset by increases in overfire air entrainment.

OFFGAS COMPOSITION

     To apply  the concepts  of "unmixedness" and  "eddy decay," it is necessary to know
the initial composition of gases entering the turbulent mixing  (overfire) region. From basic
research in kinetics of combustion  reactions, we know that oxygen and hydrogen  or carbon
monoxide react on the order of  microseconds. For all practical purposes they cannot
co-exist. However, the data  given in Table E-l show over 3% CO at times when  oxygen is
over 10%. One reasonable explanation for this co-existence is that the offgases are relatively
unmixed so that the oxygen is not in intimate contact with the combustibles.
                                        134

-------
Table E-2. SUMMARY OF BURNING RATES AND UNDERFIRE AIR CONSUMPTION

                                 Burning Rate
                                  (Ib/hr-ft2)
                                                            % of Consumed Air
 Run No. *        Heat Balance  „„     Mass Balance          Introduced Underfire

BL 22/1
J 22/1
BL 23/1
JET 23/1
BL 23/2
JET 23/2
BL 24/1
JET 24/1
JET 24/2
BL 24/2
BL 25/1
JET 25/1
JET 25/2
BL 25/2
BL 31/1
JET 31/1
JET 31/2
BL 32/1
JET 32/1
JET 32/2
BL 33/1
JET 33/1
JET 33/2
BL 34/1
JET 34/1
JET 34/2
JET 34/3
("Total Heat ~|
[3900 x 336 J
67.9
65.6
70.2
55.3
66.3
59.0
68.9
75.1
62.4
70.7
79.5
75.8
58.6
56.4
75.3
61.5
57.7
77.1
77.3
74.2
67.4
64.8
71.4
67.6
62.0
60.8
61.6
                                      rCQ2 x 23.21
                                      [ (0.51) 336 J
   Run conditions are given in Appendix A
                                            53.3                   35.5
                                            56.5                   31.3

                                            50.1                   41.5
                                            41.5                   40.6
                                            43.7                   39.3
                                            40.5                   38.5

                                            43.8                   35.7
                                            59.8                   29.6
                                            51.4                   30.6
                                            56.4                   30.6

                                            70.9                   35.7
                                            60.0                   31.5
                                            53.3                   34.9
                                            46.6                   34.9

                                            69.2                   27.04
                                            50.8                   27.04
                                            47.6                   27.04
                                                   i
                                            59.7                   28.9
                                            66.7                   30.4
                                            53.5                   28.5

                                            56.0                   29.6
                                            54.9                   29.2
                                            57.9                   29.2

                                            55.6                   33.2
                                            49.5                   30.04
                                            49.2                    23.06
                                            53.9                    30.04
                                      135

-------
     This being the case, the characterization of the offgas composition must include the
following factors:

     1 )   Composition of combustible fraction,
     2)   Fraction of combustibles vs. fraction of air, and
     3)   Degree of unmixedness between two fractions.

These are discussed separately.

COMPOSITION OF COMBUSTIBLE  FRACTION

     The data gathered from earlier measurements* of bed offgas compositions suggest that
once the oxygen is depleted, a burning refuse bed acts as a gasifier, producing CO, CO2 , H2 ,
and H2 O with minor amounts of hydrocarbons. This must occur at some place within the
bed in order to have combustibles present in the offgas. The controlling equilibrium for this
portion of the offgas is between these gasifier products.

     Tentative bed-burning models have been developed*** on the assumption that the
water gas shift reaction controls the  gasification  equilibrium.  The  experimental basis
supporting the postulated equilibrium was previously limited to data taken by Kaiser t on
the Oceanside  (N.J..)  incinerator and data taken at M.I.T. on synthetic refusett. The data
collected in this program also support this hypothesis.

     The composition of the combustible fraction can  be  determined as a function of
CO/CO2 . The calculation scheme is as follows:

For the reaction:

                       CO2 + H2 -»• CO + H2 O, (water gas shift)                  (E-l )

the equilibrium constant is defined as:
                                     (H,0)(CO)
                                                                             (E-2>
From the material balance:

                CH04(H2O)06 -0.7H2O+ l.lOj -»CO2 + 1.5 H2O,             (E-3)
                                       136

-------
the equilibrium concentration can be calculated as follows:

                     Let: C0/C02 = x, then (H2 O) = (— — J (H2 )              (E-4)


     Choosing as a basis one mole of CO2 corresponding to x moles of CO, then the total
moles of carbon is (1 + x). From the  material balance, the total hydrogen (in the form of
H2) is determined to be 1.5 (1 + x), since the ratio H2/C equals 1.5. Using the equilibrium
relation above  and noting the sum  of H2O and  H2  equals the total hydrogen  yields
1.5(1 + x); then the relative proportions of H2O and H2 can also be calculated. Finally, the
amount of N2 present is 3.76 times the amount of oxygen consumed. The latter is equal to
the total oxygen present (CO2 + 1/2 CO+ 1/2 H2) minus  the  amount  of O2 originally
contained in  the initial  (1 + x) moles of refuse, i.e., 0.65 (1 + x). These proportions are
tabulated below.

                        Relative Proportion of Combustion Products

        Product                  Mols

          C02                   1

          CO                    x

                                 1.5 (1 +x)K
          H2O
          H                      1.5 (1 + x)x
                                     I             / 1.2K_ -0.3x
          N2                     3/76    l+
-------
u>
00
                 o
                  (N











D





D

D



Q



m ,




D



Q
[
a
D
a
a a
BD
3 D
Z
AA
A







3


D
n° a

A
O
A^b^
©A A




O






o

o
OQ
) GP AA Qj>>
C

O











D
o
0


                         750

                          O ZONE 1
                          A ZONE 2
                          D ZONE 3
1000
1250             1500
   TEMPERATURE, °F
1750
2000
                                                  Figure E—2  Plot of H2 O/CO2 versus temperature

-------
         programs. The assumption is based on the concept that the mechanism for
         forming CO and  CO2  in the bed  of an incinerator is the same as that for
         gasification of other organic combustibles. This is approximately valid before
         combustible burnout begins in the overfire region.

     2.   The ratio  can  be  determined by actual sampling over the bed. The sampling
         is extremely difficult and the analysis must always be suspect.

     In  both cases,  the  ratio is only a crude estimate, but the sophistication required to
predict the exact carbon ratio is neither currently available nor particularly warranted.

     The analysis in Section V was based upon the CO/CO2 ratio as a characterization of
the extent to which the combustion reaction has gone to completion. From the summary of
the bed data in Table E-l an empirically determined ratio ranges from 0.02 to 0.6 with the
average  being approximately 0.21. These values can be compared reasonably well with
values obtained for coal gasification or for  pyrolysis of refuse. These values are probably
minimum because some overfire mixing has occurred before the gases from the bed can be
sampled in the sample probe. Thus, a certain amount of reaction would be expected within
the probe itself, reducing the amount of CO and correspondingly increasing the amount of
CO2 measured in each sample.  Hence, the empirically determined CO/CO2 ratio is  not
inconsistent with the types of initial  ratios necessary  to  develop  a condition in which
quenching would dominate  (Appendix C).

     A  second tenet of the equilibrium  model  is that  combustion is  relatively uniform
throughout the bed, i.e., drying, gasification, combustion,  burnout, etc, occur simultane-
ously. Although this is not rigorously true, the errors introduced by such  an assumption are
no greater than those resulting from analytical measurements.

     Verification of these two points: the equilibrium of gases and the uniformity through-
out the furnace is given below.

     Two independent checks were made of the water-gas shift equilibrium. For the data
collected in  the  breech during the  baseline tests, the water-gas shift constant is shown in
Figure E-3. With respect to these data,  two things are noteworthy. First of all, the ratio
H2 O/CO2 remains relatively constant so that in the breech area the ratio  CO/H2  is expected
to remain constant as well. The large amounts of water and carbon dioxide present act like a
flywheel, so that large changes in the CO and H2 concentrations are not at all seen as similar
changes in water or CO2. For refuse, the ratio H2O/CO2  is equal to 1.5 so that the water-gas
shift constant could be represented by the equation:

                                 Kwg = 1.5CO/H2.

     The fact that the calculated value of Kw g is even close to the theoretical value is a very
encouraging support of the water-gas equilibrium theory. The jet data were not used for this
because of the difficulty detecting H2 with the modified probe.

                                        139

-------
   15
  10
o>
9

8

7
     4

     3


     2

     1

     0
       700
                                  AA
              KEY:
                       = (CO)(H20)
                   Wg  " (CO,KH,)
         ^2/^2

o =  ZONE  I
n =  ZONE  II
A =  ZONE  III
                                 Kwg (THEORETICAL)
                1000
                 1250
                                                1500
                                     TEMP., °F
                                                             a
                                                             D
                  Figure E—3   Water - gas shift constant from breech data
                                                                   2000
                                      140

-------
     Additional confirmation- of the equilibrium can be seen from data taken directly over
the drying grate. In these tests in which the concentration of CO and H2 is sufficiently high
to obtain good analytical measurements, the experimental value for Kw g is remarkably close
to the theoretical value. A summary of the results of these tests was given in Table E-l. It is
also interesting to note that the ratio of hydrogen to carbon is consistent with the H2O/CO2
ratio of 1.5.

     The condition of uniformity is  also easy to substantiate  by considering a plot of
H2O/CO2  as shown in Figure E-2. If,  as the earlier models proposed, the drying zone and
the pyrolysis zone were distinct from one another, then one would expect to see a reduction
in the ratio of H2O/CO2 as the refuse travelled through the incinerator. For example,  refuse
on  the first grate would have only a small amount of combustion or pyrolysis and a large
amount of drying so that the ratio of H2O/CO2 would be larger than normal, reflecting the
large amounts of water resulting from  the drying and  small amounts of CO2 from combus-
tion.  On the other hand, the gases over the burnout grate should be completely dried and
pyrolyzed so that very small amounts of hydrogen would be expected  and the ratio of
H2O/CO2  would be correspondingly small. In contradiction to this, the data indicate that
there is no trend in H2 O/CO2, but rather that the ratio tends to be constant at approxi-
mately 1.5.  The 1.5 value is consistent with the ratio of hydrogen to carbon for a  refuse
with 25% moisture (by weight). This tendency was observed both in the breech and over the
bed and cannot, therefore, be attributed to overall mixing before the sample was taken.

     Contrary to  expectations, the H2O/CO2 ratio tends to be higher over the third grate
than  it  does over the  other two. We do not attribute this to any characteristic of the
refuse-burning process, but rather to  the fact that the hot ash  dropping into  the  water-
quench  hopper  evaporates a sufficient amount of water to cause a shift in the H2 O/CO2
ratio  over the third grate, an  effect  that would  be  witnessed in zone 3 because of the
stratification of temperature layers in  the breech of the furnace.  Indeed, a material balance
about the third grate  indicates that the water introduced by this evaporation mechanism
would be sufficiently large for such an effect to be observed.

DETERMINATION OF OXYGEN FRACTION

     Significant amounts of oxygen were present in most of the samples taken above the
fuel bed. Since,  at furnace temperatures, oxygen does not coexist with supra-equilibrium
concentrations of CO and hydrogen  for more than a few milliseconds, the detection of both
oxygen  and fuel components in  the  samples is indicative  of intermittent sampling from
fuel-rich and oxygen-rich eddies. The oxygen observed above the bed probably originates in
the underfire air which passes unreacted through blowholes or overfire air carried down by
buoyancy forces. Table E-3 presents estimates of the amount of oxygen present in the gas
samples expressed both as a fraction of the stoichiometric air requirements and a fraction of
                                        141

-------
the total air in the sample (determined from a nitrogen balance). With the exception of one
sample, the available oxygen exceeded the stoichiometric oxygen requirements. The role of
overfire air jets for the present conditions therefore should be principally to provide mixing,
and caution should be exercised that the jets do not quench the combustion reactions.

   Table E-3. AVAILABLE OXYGEN IN SAMPLES WITHDRAWN FROM ABOVE THE FIRST GRATE

                             Available O2             Unreacted 02    Unreacted O2
       Bed Test 1     	    	    	
     Probe Position   02  Required to Complete Combustion   02 Supplied    Total Offgas

           1                       2.82                   0.58
           2                     74                      0.89        0.90 >  0.75
           3                     86                      0.69

       Bed Test 2
     Probe Position

           1        ,               0.875                  0.31
           2   ,                    9.1                    0.69        0.68 >  0.46
           3                       8.2                    0.41
     A peculiarity in the data resulting from the design of the incinerator merits comment.
The unreacted fraction in Table E-3 peaks at position 2, which was directly below one of
the roof air jets. An attempt made to close the damper to the roof jets between bed tests 1
and 2 resulted in a reduction of the unreacted air. This tended to support the hypothesis
that unreacted overfire air was entrained in the sample. Because of the presence of overfire
air, it is not possible to determine, on the basis of these measurements alone, the fraction of
the underfire air  which  is unreacted.  The bed-burning  data suggests that most  of the
unreacted oxygen originates in the overfire jets.

DEGREE OF MIXING

     The  data given in Table E-l  could be used to calculate an initial unmixedness  factor
according to the procedure outlined in Appendix B. In fact,  the value  for $0  has already
been  given  in Table  E- .  The corresponding values for (A/C^)O are shown in Table E-4.
Values given for position  3, near the end of  the first grate, are  similar to the  value for
(v/c^o  calculated in Table E-3.  One might be tempted to rationalize,  explaining that
position  3 is closest to the center of the  furnace, particularly to the first burning,  and  is
therefore  more  representative  of the  bulk of combustion gases found  in  zone I of the
breech.  The data  from zone I were  used  to calculate the  values given  in Table E-3.
Unfortunately, the data are  too scattered  and inconsistent for such a rationalization; no
conclusions are apparent from the data.
                                         142

-------
                   Table E-4.  INITIAL UNMIXEDNESS FACTORS
                                                        C0/C02
0.25
0.811
0.462
0.300
0.50
0.611
0.409
0.249
1.0
0.365
0.340
0.185
Position
   1                2.3
   2                0.60
   3                0.58
Average                                  0.524             0.423            0.297

Bed 2

Position
   1               - 2.0                  1.273             2.982            4,951
   2                1.14                 0.556             0.457            0.332
   3                1.18                 0.286             0.184            0.061

                                        0.705             1.208            1.781
                                        143

-------
                                   APPENDIX F

                  DERIVATION OF THE UNMIXEDNESS FACTOR

     The following derivation is  a summary  of the development of the "unmixedness
factor" concept based upon Reference  . The variable, 3>, the ratio of oxygen in the sample
to the oxygen required to complete combustion of the sample, can be calculated as follows:

     If y is the mole fraction of offgas in the offgas/air mixture,  then the mole fraction of
air is (1 — y). Also, since 4.76 • n   moles of air react with 1 mole of offgas, then the
amount of offgas required to react with (1  - y) moles of air is [(1 - y)] /(4.76 • n02 ). The
remaining offgas after the available oxygen  has been consumed is equal to the initial moles
(y) minus the moles consumed, or in other words:
                                            4.76

Since
                           Excess offgas =  y — - -                    (F-l)
                                          L     4.76 -n02  J

                                          /
                                          \
                                                                             (F'2)
                              excess offgas = a (y - ys)                         (F-3)

where

                     a = constant (1 +4.76 -n02)/(4.76 -n^).                   (F-4)

The oxygen required to complete combustion is n0  times the amount of offgas, so that:

                                  Ot req = (C - Cs)

     Likewise, in a sample where y is less than ys, the air concentration is still (1  — y) • Ct,
but in this case oxygen remains unreacted, i.e., the limiting reactant is offgas. In this case,
y • Ct moles of offgas require (4.76 • n0,)' (y ' Ct) moles of air. The remaining air is:

                  Excess air = (1 - y) • Ct - (4.76 • n0 a ) • (y • Ct)

                            = (l+4.76-n02)(y-ys)-Ct                      (F-S)
                                        145

-------
and
                             n+4.76-n0.n
                                             •(C-C,) = a'(C-C,)           (F-6)
Note that Eq.  (F-6) is exactly the same as Eq. (F-4), except that it applies to the region
where y>ys, while Eq.  (F-4) applies to the region where y/7rn" e ~"  +erf(n)         1C
                    $ = 	*-	—	   ; n =  ^=  —^            (F-9)
                           1 + J   e   + erf(n)
    A plot of 4> vs  —— is given as Figure 3 in Section V.
                                      146

-------
Length
Volume
Mass
                             APPENDIX G
                CONVERSION FACTORS FOR SI UNITS
              1 ft.   = 0.3048 m
              1 in.   = 0.0254
           1 cu. ft.   = 0.02832 m3
              1 Ib   = 0.4536 kg
           lib/ft3   = 16.017 kg/m3
            llb/hr   = 0.126xlQ-3 kg/sec
           1 Ib/hr-ft2= 1.355 x 10"3 kg/m2-sec
         1 ton/day   = 0.0105 kg/sec
           1 ton/hr   = 0.438 x 10~3 kg/sec
Flow Rate
             1 cfm   = 0.472 xlO'3 m3/sec
            1 ft/hr  = 0.0847 x 10~3 m/sec
           1 ft/sec  = 0.3048 m/sec
Energy
            1 Btu    = 0.252 kcal
          1 Btu/lb   = 0.555 kcal/kg
          1 Btu/lb°F = 1.0 kcal/kg °C
          1 Btu/hr   = 0.070 x 10~3 kcal/sec
        1 Btu/hr ft2  = 0.754 x 10~3 kcal/m2 sec
        1 Btu/hr ft3  = 2.47  x 10~3 kcal/m3 sec

Temperature

            1°F    = 0.555°C
          Temp (°F) = 0.555 Temp (°C) + 32°F
or 1.8 [Temp (°F) - 32°F] = Temp (°C)
                                 147

-------
                                 TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
 1. REPORT NO.
 EPA-600/2-75-016
                            2.
             3. RECIPIENT'S ACCESSION>NO.
 4. TITLf AND SUBTITLE

 Incinerator Overfire Mixing Demonstration
             5. REPORT DATE
             August 1975
                                                       6. PERFORMING ORGANIZATION CODE
 7. AUTHORISL
         T.J.Lamb, R.H.Stephens, C.M.Mohr, P.L.
 Levins, L.K.Fox, and A. F.Sarofim
             8. PERFORMING ORGANIZATION REPORT NO

             ADL-73722
 I. PERFORMING ORGANIZATION NAME AND ADDRESS
 Arthur D. Little, Inc.
 20 Acorn Park
 Cambridge, MA 02140
             10. PROGRAM ELEMENT NO.

             1AB015; ROAP 21AUZ-015
             11. CONTRACT/GRANT NO.
             68-02-0204
 12. SPONSORING AGENCY NAME AND ADDRESS
                                                       13. TYPE OF REPORT AND PERIOD COVERED
 EPA, Industrial Environmental Research Laboratory,
  Research Triangle Park, NC  27711  and Municipal
  Environmental Research Laboratory, Cincinnati,
  Ohio  45268
             Final: 1971 - 1974
             14. SPONSORING AGENCY CODE
 IS. SUPPLEMENTARY NOTES
 16. ABSTRACT
 The report gives test data and conclusions from a testing program of an existing
 municipal incinerator.  Measurements were made in the furnace breech, directly
 over the refuse bed, and in the furnace stack.   Generalized models were developed to
 describe bed burning, furnace flow behavior, and overfire mixing phenomena. Design
 guidelines were  developed for operating variables, such as underfire air distribution
 and feed rate, and design features, such as mixing jet placement and furnace config-
 uration.  Jets were shown to be effective in reducing combustible emissions through
 temperature control. They were only marginally effective in inducing increased
 turbulence within the furnace. Analysis of test results and rationale for the conclu-
 sions drawn are discussed thoroughly.
 7.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.lDENTIFIERS/OPEN ENDED TERMS
                         c. COS AT I Field/Group
Air Pollution
Incinerators
Refuse Disposal
Combustion
Air Pollution Control
Stationary Sources
Overfire Mixing
13 B
                         21B
 8. DISTRIBUTION STATEMENT

 Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
      162
20. SECURITY CLASS (Thispage)
Unclassified
                         22. PRICE
EPA Form 2220-1 (9-73)
                                        149

-------